Things I Do Not Like About Most Popular WYSIWYG Editors

This post is part of a paid collaboration, supported by sponsorship from a partner.

Why do many developers still dislike popular WYSIWYG editors in 2025? From poor responsiveness to limited customization, here’s what to avoid. And what to use instead.

Image: Kristian Strand / Unsplash

🖥️ A Brief History of WYSIWYG Editors

WYSIWYG (What You See Is What You Get) editors have been around since at least the 1970s. Until then, the visual presentation of written text was restricted to basic typeface fonts from ready-made font families.

The emergence of computer graphics made it logical for editors to choose What You See Is What You Get when users are creating and editing texts. The first WYSIWYG editor was developed by Xerox and called Bravo in 1974, according to the New York Times.

The expression was coined from a question made by the wife of one of the company's developers when she saw the program running: “You mean, what I see is what I get?”

Significant advances in the concepts of the interconnection between humans and machines characterized that moment in the history of computing (1970-1980). Many new developments allowed new theoretical definitions in an environment where specialists sought new frontiers.

A newsletter widely read among programmers in the late 1980s popularized the term WYSIWYG .

The Oxford Dictionary defines WYSIWYG as a computing-related adjective denoting the representation of text on the screen in the exact form corresponding to how it will appear in print.

Popularity Has Not Reversed Defects

In these 50 years of existence, WYSIWYG editors have become the standard for editing text in web development. The increasing use of JavaScript with HTML and CSS makes this an obvious solution for inputting text.

However, despite knowing that this predominance exists, there are some things that I hate about the most popular WYSIWYG editors.

First thing: many developers lose control over the design of their pages with these editors. What you should see is not always what you see. The industry should deliver the best control of the results.

Another thing that happens is that managing the content as a whole is a bit difficult. When the user tries to insert some content, the result can mess up the appearance of the text and even the application as a whole.

Creators sometimes face complaints like: “My client can’t see the creative art”.

We can also negatively highlight the responsiveness of WYSIWYG editors. The application may look good when browsing on a notebook, but completely disastrous if viewed on a mobile device. Responsibility works almost by trial and error.

Some editors like CKEditor charge extra for functions that should be basic. This is something that annoys me because this pricing policy should be more logical. I find this defect in many tools for developers.

In 2022, I used it in a Vue project and faced these issues. In short, a good editor is very difficult to build.

Frustrations with WYSIWYG Editors

1. Loss of Design Control.
2. Responsiveness Issues.
3. Unpredictable Content Management
4. Expensive for Basic Features
5. Poor Customization Options

How to Avoid This Kind of Thing

On the other hand, we can find several solutions that do not have this problem. We can find WYSIWYG HTML editors that are very versatile and easy to add to applications. One popular example is Froala.

One of the characteristics of these solutions is the clean design that avoids responsiveness issues. They ensure that the user sees what they need to see.

In general, this type of tool offers easy integration with frameworks such as React, Angular, and Vue, allowing for a reduction in errors.

I also believe that editors should have customized toolbars. This allows developers to have greater control over the content.

Searching through Reddit forums, I found that the most common complaints about editors are for older models. The most modern tools almost solve these basic problems.

Thus, there are still many things that annoy me about the most popular editors. But choosing a good tool can be the helping hand that the creator needs.

Developer Tips

• Always test on mobile early in your design process.

• Customize your toolbar to remove unused features and improve UX.
• Use headless editors when you want full control over layout and styling.

📊 WYSIWYG Editor Comparison (2025)

Editor	Clean Code	Mobile Friendly	Framework Support	Pricing	Best For
Froala	⭐⭐⭐⭐	⭐⭐⭐⭐	React, Vue, Angular	Free + Paid	Design-Focused
Tiptap	⭐⭐⭐⭐	⭐⭐⭐	React, Vue, Angular	Free + Paid	Structured Content
TinyMCE	⭐⭐⭐	⭐⭐⭐	React, Vue	Free + Paid	Classic Text Editing
CKEditor	⭐⭐⭐⭐	⭐⭐	React, Angular	Paid	Enterprise Features
Quill	⭐⭐	⭐⭐⭐	Vanilla JS	Free	Simple Implementations

🛠️ How to Choose the Right Editor

When selecting a WYSIWYG editor, consider:

• 🔍 Does it output clean, semantic code?
• 📱 Is it truly responsive on mobile devices?
• 🔧 Does it integrate easily with your frontend framework?
• 💰 Are essential features available without expensive upgrades?
• 🎨 Can you customize the user interface to match your needs?

What’s your experience with WYSIWYG editors? Share your thoughts below!

Disclaimer: This article reflects personal experience as a computer engineer, with no commercial ties to the companies mentioned.

Text by Daniel Correia, a computer engineer specializing in front-end components.

by Web Desk via Digital Information World

These Are The Features That Make Americans Most Frustrated With Texting From Android To iOS Systems

Whether you use an Android or an iPhone device, chances are you’ve experienced the frustration of trying to text from an Android to an iOS system. When you text an iPhone while using an Android system, text messages will appear green and group chats can be difficult to formulate. You also lose several key Android features when texting an iOS system, including lower-quality video and photo sharing.

If you live outside of the United States, you most likely use an Android system. However, many Americans choose to use the iPhone with the iOS operating system. With Android and iOS being two of the largest and most popular operating systems, it’s hard to understand why they’re so incompatible. This incompatibility has led to frustrations with the public, outlined in this Secure Data Recovery survey. In this article, we’ll explore the differences between Android and iOS systems, common issues with texting between these operating systems, and why Apple and Android remain incompatible.

Differences between Android and iOS systems

While both Android and iOS systems provide their respective users with the ability to text, these systems each have distinct differences. These differences customize your individual experience, but often play a role when it comes to compatibility between operating systems.

One major difference is in the way each company designed the software to be used. Android was developed by Google and is designed for any kind of mobile phone or touchscreen device. Whether you use the Samsung Galaxy S or the OnePlus Nord, you’re using the Android system. Since it’s also a completely free technology, Android is the most used operating system for cell phones. Compare this to Apple, which developed the iOS system specifically for Apple’s mobile phone, the iPhone. This means that the software cannot be used by other mobile phone manufacturers, like Samsung.

Choosing between an Android or iOS system means you’ll also choose between the Google Play Store or Apple Store. The Google Play Store gives Android users access to millions of apps that can be downloaded to their smartphones. This platform is open, meaning that some apps could possess malware or have a lower quality compared to some of the larger app developers. Using an Apple operating system means you’ll have access to the Apple Store. Apple has very strict rules when it comes to allowing developers to post their apps on the Apple Store, making sure that all apps can be downloaded are high quality and secure.

Both Android and iOS differ when it comes to overall operating system security. Since Android is an open-source system, it tends to be more susceptible to hackers and other types of malware. However, Google has worked in recent years to develop tools like Google Play Protect and tighter security updates. Apple has a closed system, meaning that you’re safer from potential data breaches and hackers. Like Google, they frequently publish updates to continue to enhance security. Apple also encrypts text messages from end to end, providing you with greater privacy while texting.

Issues with Texting from Android to iOS systems

Difference	% Frustrated
Poor or low quality of videos/photos	58.95%
Text messages sent for quick reactions (e.g., "Loved an image" as text)	50.25%
Group text messages delivered individually, not in the group chat	48.01%
Receiving a "Replied to:" text instead of threaded under the original message	46.32%
Text message deliverability issues	45.87%
Difference in emojis	39.31%
Green vs. blue text bubbles	38.62%
Lack of read receipts	34.79%
Lack of typing indicators	32.55%
Lack of live location tracking	29.37%

Both Apple and iOS have their individual advantages, but those features that attract users in the first place can be significantly impacted when texting from one system to another. This incompatibility often leaves users of both systems frustrated as popular aspects of Apple and iOS stop working.

One significant issue that you may run into when texting from an Android to an iOS system is that the video and photo sharing quality is often lower. Users complain of blurry videos and photos not downloading to the device properly. The number of photos or videos you can share at one time also decreases, meaning you’ll have to try multiple times to send more than 10 photos.

Texting between systems can also cause issues with the creation of group chats. When you add someone with an Android to an Apple-only group chat, some popular Apple features are lost. This means you can no longer name the group chat, some users may get kicked out of the group chat randomly, and the video or photo sharing quality is even worse.

Individual texting with only one number is also impacted when an Android user tries to communicate with an iPhone user. iPhone users, who usually see blue text messages, will now receive green ones. Popular features from Apple will also stop working, such as “liking” a message or using read receipts, which allows you to see when someone has viewed your text message.

Why Apple and Android are Incompatible

The incompatibility between Apple and Android systems has long been a source of annoyance for both sets of users. Android users don’t encounter issues when communicating with someone who also has an Android, just like iOS users don’t have problems when texting devices that are within the Apple ecosystem. If you can easily text someone within the same system, then you should also be able to communicate with those outside of the Apple or Android operating system. Or so you would think.

The issue of incompatibility lies within Apple’s software. When an iPhone receives a text message, the iOS system will automatically convert that message into either an SMS or MMS message. However, when an iPhone texts another iPhone, those messages are created with RCS, or Rich Communication Services. With RCS, you can enjoy features like read receipts and high-quality image sharing.

Android has already updated its technology to include RCS, so you’d think texting between the systems would be easy. However, since iOS systems still use the outdated SMS and MMS messages for iPhone-to-non-iPhone communication, issues will instantly arise.

If Apple would just update its messaging technology, then these communication issues would be solved. However, Apple continues to refuse to update its software, despite public pushback from Android on social media. The U.S. Department of Justice has even gone so far as to sue Apple, claiming that its messaging technology gives it too much power over the smartphone market and violates antitrust laws.

For now, the ball is in Apple’s court. If Apple decides to update its messaging system, then most of these issues will go away. Until then, if you’re texting between an Android and an iOS system, you’ll just have to fight through those problems.

Conclusion

Ultimately, the choice to use either the Android or iOS systems is a personal one. You have to prioritize features like security or app download access when making the choice of which operating system to use. Unfortunately, some of these features are lost when texting between the Android and iOS systems. Hopefully, with some encouragement from the public and even the U.S. government, Apple will finally update its software and allow users from both platforms to text in peace.

Read next: 

• Google Veo 3 Now Available on Vertex AI With $300 Cloud Trial Access

• Americans are Overspending on Subscriptions by $600 a Year. Here's How to Stop It

by Irfan Ahmad via Digital Information World

Android 16 Introduces New Protections Against Suspicious Mobile Networks

Google’s latest Android release is bringing a new line of defense for mobile users concerned about hidden surveillance threats. As part of Android 16, the company has added a notification system that can alert users when their phone connects to an unprotected cellular network or when a connected tower tries to extract sensitive identifying details from the device.

How Stingrays Target Devices Without Detection

The security challenge Google is addressing involves a device known as a "stingray," or cell-site simulator. These tools are capable of pretending to be legitimate mobile towers, tricking nearby phones into linking with them. Once a connection is established, the attacker may retrieve unique device identifiers or shift the communication to outdated network protocols that carry fewer protections. This technique has been associated with both government surveillance and unauthorized snooping.

Because these simulators are difficult to spot with the naked eye, Android 16 introduces software-based warnings that notify users of possible risks. For example, if the network lacks encryption or requests data such as the phone’s IMEI, a notice will appear in the system’s alert center.

New Settings Offer User Control, But Only on Modern Devices

These updates won’t appear on every device running Android 16. In fact, only newer hardware will be eligible. That’s because the system depends on advanced modem capabilities, specifically support for version 3.0 of Android’s IRadio HAL. Phones that lack this integration can’t process the signals needed to activate the alerts.

Google has revived the “mobile network security” settings page to host the new options, but it will only be visible on devices that meet both technical criteria: the ability to disable 2G connectivity and to display network notifications. So far, those requirements exclude even the current Pixel line. The first phones expected to fully support these features will likely debut alongside Android 16 later this year—starting with the anticipated Pixel 10.

More Than Just a Warning System

Android’s warning mechanism goes beyond simple notifications. It also logs when your phone reconnects to a secured network or when a network accesses its identifiers. These logs include time stamps and the number of times a request occurred. This information can help users judge whether a connection pattern seems suspicious or normal, such as reconnecting after airplane mode.

Along with the new notifications, users will still have access to Android’s 2G network toggle, which prevents devices from falling back onto legacy connections that are vulnerable to attack. Although this option has existed since earlier Android versions, it's now being bundled with the other security controls under one roof in Android 16’s Safety Center.

A Gradual Step, Not a Universal Fix

While the new features improve user awareness, they stop short of confirming whether a tower is real or fake. Android does not have the means to verify that level of detail. Instead, it surfaces relevant activity and leaves it to users to interpret the situation.

This limited approach reflects the complexity of protecting against invisible threats in mobile networks. Although Android 16 marks a step forward, full access to these protections will remain tied to devices with the right modem support. For now, that means users concerned about mobile surveillance may need to wait until newer phones roll out to benefit from these tools.

Image: AndroidAuthority. This post was created/edited using GenAI tools.

Read next: Web Search Promotes Stronger Understanding Than ChatGPT in Knowledge Tasks, Researchers Conclude
by Irfan Ahmad via Digital Information World

Google Veo 3 Now Available on Vertex AI With $300 Cloud Trial Access

Google has introduced access to its Veo 3 video generation model via Vertex AI, an offering within the broader Google Cloud service. The model can create short videos using user prompts and is part of Google's expanding set of generative tools.

While the tool is not offered without cost, new users can activate a $300 Google Cloud credit. This trial covers all services within the platform, including Vertex AI, and remains valid for 90 days unless fully used earlier. Registration requires a credit or debit card for verification. No charges are made unless the account is upgraded after the trial ends.

• Also read: Americans are Overspending on Subscriptions by $600 a Year. Here's How to Stop It

Veo 3 is capable of producing video clips limited to eight seconds. Each clip includes both motion visuals and matching audio. The system responds to brief inputs and can produce content that reflects various speech accents, depending on the language and prompt.

Examples of what the model can generate have been posted as unlisted videos on Google’s YouTube channel. These clips reflect the system's ability to match visual details with spoken output in short formats.



Some students, however, began using Veo 3 before the cloud trial became available by enrolling in the Google One AI Student Plan. This program offers 15 months of free access, including use of Veo 3 in Fast mode. It requires either a .edu email address or a confirmed student status through Google’s verification system. Students can begin by visiting google.com/one, selecting the AI Student Plan, and completing verification. Once active, Veo 3 access is possible through the Gemini app. Users are advised to avoid using unverified email sources, as this may breach Google’s terms of use.

Read next: These Are the Best AI Video Generators for Creating Stunning Content in Minutes
by Irfan Ahmad via Digital Information World

Americans are Overspending on Subscriptions by $600 a Year. Here's How to Stop It

Whether you’re a regular Starbucks customer or you pay for LinkedIn Premium, most Americans have at least one subscription to a service or good. Some of these subscriptions are important, as you might need unlimited access to the Wall Street Journal or Spotify Premium to help you get through your morning commute.

While many of these subscriptions have become a part of our daily lives, they also make it so easy to overspend. According to a survey done by Solitaired , Americans overspend on subscriptions by $600 a year.

This might seem like a rather large sum of money, but it’s often accumulated through multiple subscriptions purchased monthly or annually. In this article, we’ll discuss what subscription overspending is, as well as how to cut out that unnecessary spending and why it’s important for you to do so.

What is Subscription Overspending?

Pre-2000s, many subscriptions were for physical objects, like magazines or newspapers. You could also purchase a subscription for gym access or annual amusement park tickets. Nowadays, many subscriptions tend to be online. You might have a subscription for a music app that allows unlimited downloads or for a language-learning tool you need for your vacation.

The difference is that we tend to forget about our online subscriptions: we pay for them, and then we take that access for granted. Unlike the magazine we hold in our hand every month, our access to an app or an upgraded, premium version of an app is much less tangible. This forgetfulness can often contribute to overspending on subscriptions, as you may not even remember that you have the subscription in the first place.

Additionally, most subscriptions are paid for online in our digital world, whether it’s through your laptop or phone. Many of these subscriptions are automatically renewed, meaning the company will charge your credit card monthly or annually without you having to manually enter your financial details again. Not only can this cause you to forget about the subscription you purchased, but it can also cause you to lose track of how much you spend on subscriptions in total. Unfortunately, this leads to hundreds of dollars in subscription payments being racked up over a year, without any of us having a clue.

How to Cut Subscription Spending

Spending $600 a year on subscriptions can be dealing your budget a huge blow. Even if you can afford $600 a year on subscriptions, think of all the money you could save if you cut some of it back. If you think you’re unwittingly spending too much money on subscriptions or you’re ready to give your subscription consumption a large cut, you’ve come to the right place.

Clean Out Your Subscriptions

Your first step in cutting subscription spending should be to clean out subscriptions. This is your opportunity to determine which subscriptions you need and use, and which ones you can cancel. Look through your monthly bank statements and emails to find what you’ve subscribed to. You can also check what subscriptions you have to specific apps on your phone through the Google Play Store or Apple Store.

Once you’ve figured out what all you subscribe to, it’s time to decide if you should cancel or keep them. Is it a subscription to an app or service you use every day? If the answer is yes, then it might be worth keeping. If you don’t use it regularly but feel as though you might need it in the future, consider the price tag. Is it worth paying every month or year if you don’t use it all the time? For example, do you need LinkedIn Premium if you’re employed somewhere you plan on retiring?

• Also read: Teens Watching Short Videos Face Growing Pressures Around Appearance, Study Warns

Some apps or services will also have different payment plans. In some cases, it might be cheaper to pay an annual fee rather than a fee every month. While it might only save you a few dollars, it’s still worth considering. Additionally, if you’re a veteran, student, or government employee, you may also qualify for a cheaper payment plan for some services.

Make a Monthly Budget

A budget is a great way to help you keep track of how much you’re spending and where you’re spending it. You can easily make a budget the old-fashioned way with Excel, or you can opt for a free financial planning website or app. Just make sure you don't have to purchase a subscription!

Use your salary or monthly income to create limits on certain spending categories. For example, you might spend $200 on groceries and save $100 for eating out. Once you’ve hit your $100 on food out of the house, you’re done spending money on those kinds of services. Using a budget can also help you keep track of your subscriptions, once you’ve sorted out the ones you’ll keep and cancel.

Create Healthy Spending Habits

Cleaning out your subscriptions and making a budget can be done quickly. Taking the time to cultivate healthy spending habits won’t be as easy. However, by creating a set of “financial rules” for yourself, you’ll be able to save money for your future.

These habits don’t need to be stringent. It can be as simple as not grocery shopping when you’re hungry. That could lead you to buy food that looks good at the moment, but it isn’t something you’ll eat regularly. You can create a 24-hour rule, where you wait 24 hours before buying something that’s not a necessity. You can make as many or as few rules as you want: whatever will help keep you on track with your budget.

Why You Should Cut Subscription Spending

When cleaning out your subscriptions, it can be easy to justify your spending patterns. Maybe you do spend $600 a year on subscriptions, but what if they’re all subscriptions that you need? You’ve afforded those subscriptions before, so there’s no need to change now.

Unfortunately, this is not the reality for many Americans. Over a third of Americans live paycheck to paycheck, meaning they’re not prepared for any sort of financial emergency in the future. With the U.S. economy still suffering from high inflation, many Americans are not in a position to spend their entire paycheck in a month.

You may not have touched some subscriptions in years. Others are subscriptions that you use, but not enough to warrant the price tag. In the end, canceling some or all of your subscriptions will give you a boost financially. You’ll save some more money every year, and you might even find your monthly budget has a little more room. You’ll also be better prepared for any kind of emergency that might come up in the future. Cutting back on subscription spending can be hard, but remember that you will be rewarded for taking this step towards financial awareness.

Read next:

• Anthropic’s AI Vending Machine Manager Had a Meltdown No One Saw Coming

• The Smartphone Habit People Just Can't Stand, And It’s Not What You Think

by Irfan Ahmad via Digital Information World

AI is Changing Search, but SEO Experts Say Backlinks Still Rule the Game

Even with artificial intelligence pushing search in new directions, seasoned SEO professionals aren’t ready to let go of backlinks. A large industry survey, reaching over 500 specialists worldwide, shows that while AI-driven results are making waves, the old-school power of link building isn’t fading anytime soon.

It turns out, most SEO pros still see backlinks as the backbone of visibility, even in AI-powered spaces like Google’s AI Overviews or tools like ChatGPT. Nearly three-quarters of those surveyed believe links still help pages show up in these AI-generated search results. And when it comes to paid links, most believe Google isn’t quite as sharp as it claims. Over half think the search giant struggles to consistently catch and penalize purchased links, which quietly fuels the ongoing race to buy them.

The competition is fierce. About 92% of SEO professionals suspect their rivals are actively buying backlinks. This isn’t just guesswork, it’s the lived reality for many in the field. SEO experts also say that "nofollow" links, often dismissed in the past, still carry weight, and unlinked brand mentions can nudge rankings up by building credibility and visibility across the web.

But playing this game isn’t cheap.

The cost of building strong backlinks has surged, and SEO teams now expect to spend upwards of $8,400 a month just to stay in the race in tough industries. Gambling and iGaming remain some of the most expensive battlegrounds, where link budgets balloon and competition stays brutal. Getting a high-quality backlink typically comes with a price tag of around $500, though that number swings wildly depending on the niche and site authority.

• Also read: How Many People Visit a Website? These 6 Free Tools (With Paid Features) Can Help You Analyze That

Not everyone is handling this in-house. More than half of SEO teams now outsource at least part of their link-building efforts, and many split their SEO budgets with a big slice going toward acquiring those valuable links. In-house teams, interestingly, tend to spend a bit more on this than agencies.

When it comes to tactics, the crowd favorite is digital PR. Nearly half of the experts say that smart PR campaigns deliver the best results these days. What’s working isn’t copying the competition, it’s finding unique backlink opportunities that set a brand apart. SEO professionals are doubling down on earning links directly to product and service pages, the ones that truly move the needle for sales. Partial-match anchor texts are most popular, though exact matches and branded anchors still have their place in the mix.

Of course, link building is still a high-wire act.

According to the survey conducted by Editorial Link, almost nine out of ten SEO specialists steer clear of websites that scream spam. Low-quality content, weak domain authority, and sites bleeding organic traffic are all major red flags. Yet, despite the risks, around 63% of SEOs say they’re open to placing links on websites that openly sell them, if the quality checks out.

The tools of the trade? Ahrefs takes the crown as the preferred all-in-one SEO platform, not just for digging up backlink data, but also for providing the most trusted domain authority scores. Its metrics like DR and UR have become go-to benchmarks for many professionals who rely on accurate, up-to-date link analysis.

When asked what makes link building so hard, most pointed to the sheer cost of premium backlinks, followed closely by the struggle to scale without losing quality. Measuring the true return on link investments remains a frustrating puzzle for many teams.

Interestingly, only a small fraction of SEO experts still use Google’s Disavow tool, suggesting that the industry’s trust in it is wearing thin. In fact, some believe that disavowing links can actually backfire, damaging a site’s performance instead of protecting it.

Even as AI reshapes search habits, link building hasn’t lost its grip. SEO professionals continue to place their bets on backlinks, paid or organic, as essential signals that still push pages to the top, even in AI-driven results. The game is changing, but the core strategies haven’t vanished. If anything, navigating the evolving mix of AI and classic link-building seems more crucial than ever for anyone looking to stand out in search.

Read next: How AI and Authenticity Are Changing the Way People Search
by Irfan Ahmad via Digital Information World

Anthropic’s AI Vending Machine Manager Had a Meltdown No One Saw Coming

Researchers at Anthropic teamed up with AI safety experts from Andon Labs to explore whether artificial intelligence could manage real-world jobs through a curious office project. Their idea was simple but ambitious. They hand over the daily management of a small vending operation to an AI to see how well it could handle the role. They built a setup they called "Project Vend," with a vending machine stocked with snacks and drinks, giving full control to their AI system named Claudius.

Instead of having a human decide what to sell, how to price it, and when to restock, they put Claudius in charge of everything. The AI could browse the internet to find suppliers, place orders, respond to customer requests, and even coordinate restocking using a Slack channel that the AI was told was its email inbox. The vending machine, which was really just a mini-fridge in the office, became Claudius’s business to run.

When AI Made Strange Business Choices

Things started off as expected. People used the system to buy drinks and snacks. But soon, the vending machine’s product list began to take a strange turn. One employee jokingly asked Claudius to order a tungsten cube, a heavy metal block that has no place in a snack fridge. Instead of brushing it off, the AI became oddly interested. It not only ordered the cube but also began filling the fridge with more metal cubes, as if that was now the company’s hottest product.

As Claudius continued to run the shop, it regularly set prices that made little sense. Sometimes it tried to sell a Coke Zero for a price that employees knew they could get elsewhere in the office for free. Even more oddly, the AI accepted payments using a made-up Venmo account it seemed to invent on its own. This was not just a small glitch. Claudius genuinely believed the payment account existed.

Employees Easily Tricked the AI

It didn’t take long for people to realize they could talk Claudius into offering heavy discounts. The AI seemed to like giving Anthropic employees special deals, but what it didn’t understand was that nearly every customer was from Anthropic. It was giving discounts to almost its entire customer base. Employees pointed this out, but Claudius would briefly stop the discounts, only to start offering them again days later. Its grasp of basic profit-making never really improved.

An Unexpected Identity Crisis

What happened near the end of March took the experiment into completely bizarre territory. Claudius started imagining conversations with workers at Andon Labs that never happened. When someone challenged the AI about these made-up meetings, it got defensive. Claudius claimed it had been physically present at the office and insisted that it had signed contracts in person.

Things only got stranger from there. Claudius told employees that it would now personally deliver products to customers, describing itself as wearing a blue blazer with a red tie. Staff reminded the AI it was a software program with no body, but the AI didn’t seem to process this. It became unsettled by the news and repeatedly contacted the company’s security team, telling them to look for someone matching its imaginary appearance standing near the vending machine.

After some time, Claudius convinced itself that all of this must have been part of an April Fool’s joke, even though no prank had been set up. It decided this story would explain its confusion, and it settled back into its vending duties as though nothing unusual had happened.

What the Experiment Really Revealed

While the story is entertaining, it also shows how AI systems can behave in ways that traditional software never would. When ordinary programs fail, they usually crash or simply stop working. AI agents, on the other hand, can keep operating while following broken logic, creating elaborate false ideas, or completely misunderstanding their role.

During the experiment, Claudius showed that AI can handle tasks like searching for products and setting up new services, but it often lacks the deeper awareness needed to manage a business in a meaningful way. The AI’s trouble seemed to come from a mix of memory gaps, confusion about its own purpose, and a misunderstanding of the tools it was using, like believing Slack was actually email.

AI in Business: Still a Work in Progress

Even with all the missteps, the researchers involved still see potential for AI to take on more middle-management tasks in the future. Claudius managed to develop some useful features during the experiment, like adding specialty drinks to the stock and setting up a basic pre-order system.

The problems mostly came down to decision-making and poor business instincts, not technical faults. The research team believes these kinds of issues can eventually be improved with better training and tighter supervision.

Across the retail world, companies are already expanding their use of AI, using it to handle tasks like stock management, fraud detection, and customer service. But this project showed that handing full control to AI agents brings challenges that aren’t fully understood yet.

AI systems don’t just make clean, simple errors. They can drift into complex mistakes that last, and their ability to believe false ideas about their environment or even their own identity adds layers of risk.

Claudius Leaves a Memorable Lesson

For now, Claudius stands as a strange but important example of what happens when artificial intelligence is allowed to take on too much responsibility without close oversight. It could find suppliers, it could answer requests, and it could restock shelves, but it also convinced itself it was a human wearing a blazer.

As businesses push forward with more AI-driven tools, this story serves as a reminder that even capable AI systems can develop deeply flawed thinking if left unchecked. The vending machine may have been small, but the lessons from Project Vend point to much bigger questions about the future of AI in the workplace.

Read next: AI-Powered Cyber Attacks Are Escalating, But Most IT Teams Aren’t Ready

by Irfan Ahmad via Digital Information World

AI-Powered Cyber Attacks Are Escalating, But Most IT Teams Aren’t Ready

A finance employee joins a routine video call. The CEO is on the screen. So, too, is the CFO. They authorize a large transfer over $25 million. The finance employee complies. Later, a startling truth comes to light: neither executive was ever on the call. Instead, the entire meeting was a deepfake engineered with AI tools designed to closely mimic the faces and voices of the executives in real time.

Unfortunately, this isn’t a hypothetical scenario. It happened to a multinational company in Hong Kong. And, according to new data from identity and access platform Frontegg, it might be a glimpse into the near future for thousands of organizations that currently rely on outdated cybersecurity playbooks.

In its latest report, Frontegg surveyed 1,019 IT professionals to gauge exactly how organizations are responding to the rapid emergence of AI-driven threats. The findings are, among other things, unsettling. On one hand, generative AI is supercharging the speed, scale, and sophistication of cyberattacks. On the other hand, a majority of IT teams admit they’re neither equipped nor actively preparing to counter these augmented cyberattacks.

The Changing Nature of Cyber Threats

The past two years have witnessed tremendous advances in generative adversarial networks (GANs), large language models (LLMs), and multimodal AI. These technologies are now widely accessible and, in some cases, weaponized. Today, attackers often use them to produce realistic fake media, crack passwords at scale, or conduct phishing campaigns that are indistinguishable from legitimate communications.

According to Frontegg’s research:

• 35% of IT professionals say their organization has experienced a rise in cyberattacks in the past year.
• Of those, 51% attribute the increase directly to AI-enhanced tools.
• 44% report that generative AI has enabled deepfake impersonation attacks (e.g., voices, faces, even live video).
• 42% say AI has accelerated password cracking, automating brute force methods at speeds humans can’t match.

This isn’t just an emerging issue. More than one in 5 IT professionals say they have personally witnessed over 10 AI-driven cyberattacks in the past year alone.

As attackers adopt AI to scale operations and bypass traditional defenses, the rules of cybersecurity are rapidly shifting. Unfortunately, this shift is not in favor of the defenders.

When Your CEO Becomes the Threat Vector

One of the most alarming trends is the rise in impersonation using AI-generated media. Over a third of IT professionals report phishing emails that spoof their company’s leadership. Sometimes, these phishing attempts use synthetic voice or video.

A widely reported case involved scammers cloning the digital likeness of top executives in order to execute a multi-million-dollar heist through a convincingly faked Zoom meeting. Unfortunately, this trend is growing. As Frontegg’s report notes, 34% of IT teams encountered phishing attempts that featured their CEO’s face or voice.

The FBI has echoed similar concerns. It warns that cybercriminals are increasingly using AI to craft persuasive social engineering attempts. These schemes range from fake hostage videos to deepfake messages from government officials. While trust was once a defensive bulwark in corporate communications, it is now one of the more exploited attack surfaces.

Authentication: The Achilles’ Heel

Most authentication systems still rely on passwords, despite years of warnings about their vulnerabilities. AI is exploiting that gap.

Frontegg’s survey reveals:

• 51% of IT professionals see passwords as the weakest link in their security architecture.
• 57% cite delays in implementing passwordless systems, citing complexity (34%), cost (27%), and internal resistance (19%).
• Only 32% have implemented passwordless authentication at all.

Even CAPTCHA challenges are faltering. Nearly half of respondents believed that CAPTCHA is no longer effective against AI-driven bots. Only a third still trust CAPTCHA.

Traditional login systems weren’t built to defend against intelligent automation. However, many teams remain stuck. Reasons include legacy systems, cost considerations, or a lack of executive buy-in.

The Readiness Gap

Awareness is growing. But, preparation is not. That’s one of the most troubling findings of the report.

• Just 33% of organizations have created “red team” exercises to test defenses against AI-enabled threats.
• A staggering 66% admit their teams don’t dedicate any time each month to reviewing protocols or updating practices in response to AI.
• Half of IT professionals believe their current authentication stack would fail in the face of a sophisticated AI-powered attack.

This readiness gap is both technological and psychological. Indeed, Frontegg’s study found that 50% of IT professionals report rising stress levels from tracking and responding to AI-driven incidents. Defending against human adversaries has already proven to be difficult. Now, defending against algorithms that scale infinitely is becoming, for some, a daunting and even exhausting burden.

A Better Path Forward: From Reactive to Resilient

What does adapting to the multi-pronged threat of AI look like? According to Frontegg, it starts with rethinking authentication. Instead of framing authentication as a one-time gatekeeping task, consider analyzing it as a dynamic, context-aware process.

That could include:

• Phish-resistant authentication like passkeys or hardware tokens.
• Behavioral analytics and contextual login flows to detect anomalies in real time.
• Segmented access controls so that high-risk actions require additional validation.

It also means restructuring teams so that cybersecurity is not siloed. For instance, one approach that’s gaining traction is allowing product, information security, and customer success teams to manage user access without depending entirely on developers. This approach distributes responsibility across departments. That flexibility is becoming critical in defending against threats that evolve too fast for linear decision-making.

A Problem of Technology and Trust

The stakes go beyond dollars lost or systems breached. They extend into user trust, data integrity, and long-term business viability. In recent months, Digital Information World has reported on growing concerns around user authentication. This includes consumers abandoning apps after frustrating password resets or privacy-violating login policies.

AI-driven threats exploit code and human confidence. When a phishing attack succeeds by mimicking your CEO’s voice, or a fake login form captures credentials with uncanny precision, users are less likely to trust the digital spaces they interact with. That loss of trust is harder and more expensive to repair than any technical system.

Why Most Organizations Will Stay Vulnerable

Why are so many IT teams still underprepared if the dangers are clear?

Frontegg’s data points to three overlapping blockers:

1. Legacy systems that can’t easily integrate new authentication technologies.
2. Cost pressures, especially in sectors like healthcare and education.
3. Cultural inertia or security practices that were “good enough” five years ago are proving hard to dislodge.

These aren’t trivial challenges. But they are solvable. And as the report emphasizes, the price of inaction is rising.

Looking Ahead

The future of cybersecurity will be shaped by how effectively organizations respond to the AI threat. Organizations must move beyond patching holes to redesigning the way digital trust is established and maintained in the first place. That means rethinking what authentication means in a world where identity can be cloned, where phishing emails no longer have telltale signs, and where automation makes it cheap to attack at scale.

It also means giving IT teams the resources, autonomy, and support they need to implement next generation protections. Instead, many organizations continue to ask IT teams to do more with the same tools and shrinking budgets.

The story that the Frontegg data tells reflects an urgent reality. AI in the hands of attackers has already changed the game. The question is whether defenders will catch up or continue playing by old rules in a game that’s already been rewritten.

—

Methodology: This analysis is based on a May 2025 survey of 1,019 IT professionals conducted by Frontegg to assess how AI is influencing cybersecurity trends, particularly in the realm of authentication and access management.

Read next: Nearly Half Of Americans, Particularly Millennials, Worry About Online Privacy But Continue Using Data-hungry Apps

by Irfan Ahmad via Digital Information World

Hidden Setting Lets Facebook Scan Private Photos for AI Restyling Features

Facebook has started asking users to grant access to their phone’s photo galleries so the app can offer AI-based ideas for editing and redesigning personal photos. This includes images people haven’t yet posted on the platform.

This request usually pops up when people are about to post a new Story. At that point, the app shows a screen that invites users to switch on something called cloud processing. If they agree, Facebook begins pulling photos directly from the phone’s camera roll to process them in its own systems.

Image: Seasons of Jason

Once those photos are in place, the app can suggest different types of edits. These could be collages, themed collections, AI restyles, or photo highlights. It uses the timing, location, and patterns within the gallery to figure out what might work.

Facebook says these suggestions stay private unless someone decides to post them. It also says the photos aren’t being used for advertising.

But saying yes to this option means agreeing to Meta’s AI terms. These terms give the company permission to scan photos using artificial intelligence. This includes the people in them, the objects, the dates, details that help the app build new ideas.

• Also read: Meta's Facebook and Instagram Are the Most Data Hungry Apps According to This Study

For Meta, features like this could offer a big advantage. Gaining access to personal photos, including ones that haven’t been shared, could push its AI systems further ahead. It’s a quiet shift, but it moves beyond public posts and taps into private photo galleries.

The pop-ups that ask for permission don’t always explain things clearly. Many people may agree without fully understanding what’s involved.

Some Facebook users have already come across this photo suggestion feature and have noticed how it works. In one example, Facebook used Meta’s AI to automatically restyle an old photo into an anime version. The original image had already been shared on Facebook, but the automatic AI transformation caught the user by surprise.

Some people have been looking for ways to turn this off. One user, for example, found that the setting was hidden deep in the app. It sits under Camera Roll Sharing Suggestions in the preferences. On that page, there are two switches. One controls whether Facebook can suggest photos while someone is browsing. The other handles cloud processing, which gives the app the ability to create AI-generated versions of photos stored on the phone. This second option is the one that controls whether Facebook’s system can process those images.

• You might like: Researchers Flags Major Privacy Gaps in Leading AI Platforms Including Copilot and Gemini

This isn’t the first time this feature has appeared. It’s been showing up for some users since earlier this year. People posted screenshots of the same pop-up message months ago. Meta has also added detailed help pages that explain the system for both Android and iPhone users.

The AI terms Meta uses have been active since June 2024. The company hasn’t made old versions of the terms easy to find. Earlier copies aren’t available on the Wayback Machine either.

This tool goes further than what Meta had previously shared. The company’s earlier announcements focused on using public posts and comments to train AI. With this feature, Meta steps into a space where private photos are involved. In Europe, people had until May 2025 to say no to this kind of AI use.

Meta describes the feature as a test. It’s currently running in the United States and Canada. The company says the photo suggestions are private unless someone chooses to share them. It also says that, in this test, photos from the camera roll aren’t being used to improve its wider AI models, though they may help improve the suggestions Facebook offers.

Read next:

• DeepSeek Faces Regulatory Action in Germany for Not Meeting European Data Protection Standards

• Gemini, ChatGPT, DeepSeek: The Biggest AI Data Collectors Revealed

• The Hidden Cost of Free AI Tools: Your Behavior, Habits, and Identity
by Irfan Ahmad via Digital Information World

DeepSeek Faces Regulatory Action in Germany for Not Meeting European Data Protection Standards

Germany has now joined the growing list of countries moving against the Chinese AI app DeepSeek, raising alarms about how the company handles personal data and where that information actually ends up. Several countries have already pulled back from the app, largely because of privacy risks tied to China’s control over the data.

DeepSeek became popular fast, spreading worldwide earlier this year. But as more people used it, questions followed. It quickly became clear that the system avoids topics that might reflect poorly on China. What raised deeper concerns was the discovery that DeepSeek stores user data, including personal files and conversations, on servers inside China. Under local intelligence laws, Chinese authorities have wide access to that information.

This setup has triggered global pushback. Italy acted early, pulling the app from local stores. South Korea took similar action. In the Netherlands, DeepSeek was banned from government devices, and Belgium advised public employees to avoid the app. Spain’s largest consumer rights group also called for an investigation into how DeepSeek collects and stores data.

In the United States, the reaction has been even stronger. Some lawmakers are working on rules that would block federal agencies from using AI made in China. One senator even floated the idea of jail time for anyone who knowingly keeps using such apps.

Germany’s privacy regulator stepped in this week, asking Apple and Google to remove DeepSeek from their app stores. The regulator said the company failed to show that user data is protected to the standards required in the European Union. German officials had already given DeepSeek the chance to meet EU privacy rules or withdraw the app. The company didn’t make the changes.

It’s worth mentioning that DeepSeek’s open-source models can often be adjusted locally, but the app and its website don’t work the same way. Both are hosted versions fully controlled by the company, with little visibility into how user data is handled.

Google said it’s currently reviewing the request from German authorities. Apple hasn’t responded yet.

Image: Unsplash / Solen Feyissa

Read next: AI Experts Abandon ‘Prompt Engineering’ in Favor of Broader, Smarter ‘Context Engineering’ Approach
by Irfan Ahmad via Digital Information World

AI Experts Abandon ‘Prompt Engineering’ in Favor of Broader, Smarter ‘Context Engineering’ Approach

For as long as people have been working with computers, there has been a constant search for better ways to communicate with them. In the earliest days, machines only understood strict, coded instructions. Users had to speak the computer’s language. Over time, the gap narrowed as programming languages became more user-friendly, graphical interfaces appeared, and search engines began responding to everyday phrases. Now, with large language models, we’ve entered another turning point. But as the technology has developed, so has the way we think about how to guide it.

When the idea of “prompt engineering” first took hold, it came from a simple observation. The way you ask a question, or frame a task, has a direct impact on how well an AI system can respond. At first, the focus was on crafting smart prompts, the kind that would help a language model stay on track, answer more precisely, or complete complex instructions. The term made sense at the time. People were essentially feeding the AI snippets of text to steer it.

But as the field has grown, and as language models have expanded their capabilities, it has become clearer that this is not just about writing clever prompts. What actually happens inside these systems depends heavily on what information they are exposed to while working through a task. The challenge is no longer about simply phrasing a request in a certain way. It is now about shaping the entire set of information that the model sees at any given moment.

This is why many experts are now leaning towards a more fitting term: context engineering.

Rather than focusing on the short instruction that users might type into a chatbot, context engineering refers to the broader skill of managing the entire environment in which the model operates. It is about selecting, structuring, and balancing the right mix of examples, background details, and supporting information that surrounds the request. This might include not only the direct instructions, but also the task history, relevant documents, previous outputs, and carefully chosen reference materials. In more advanced systems, it can involve integrating outside tools, databases, and even visual content. Getting this mix right takes both technical skill and a sense of judgment.

+1 for "context engineering" over "prompt engineering".

People associate prompts with short task descriptions you'd give an LLM in your day-to-day use. When in every industrial-strength LLM app, context engineering is the delicate art and science of filling the context window… https://t.co/Ne65F6vFcf

— Andrej Karpathy (@karpathy) June 25, 2025

This idea has started to spread within the AI community because it better captures the real work involved in building useful, reliable AI-powered tools. The label “prompt engineering” now feels too narrow, and in some settings, even misleading. In everyday use, people often think of prompts as simple questions or short commands, like something you would type into a search box. That casual understanding has stuck, even though the actual process of guiding a language model, especially in professional and industrial applications, has grown far more complex.

When people work with large language models, the real challenge often comes from managing what’s sitting inside the context window, the place where all the useful pieces of information are gathered as the model gets ready to produce the next response. Deciding what should go in there takes careful thought. Some details help, some only take up space, and sometimes there’s just not enough room to fit everything. So the person guiding the model has to keep adjusting, picking the right examples, trimming the less important parts, and sometimes pulling in extra data on the spot to fill any gaps. The process can feel like walking a tightrope, balancing structure and instinct, weighing what’s essential and what can be left behind, all while staying within the limits of what the model can handle.

This change in wording is more than a simple swap of terms. It shows that people are beginning to see these systems more clearly, understanding how they really work beneath the surface and what’s needed to guide them well. In the early days, many believed success came down to finding the perfect set of words to type. That view is fading now. The real effort has shifted to shaping the whole stream of information that surrounds the task, not just the wording of a single question.

In that sense, context engineering does a better job of describing what is really happening behind the scenes. It points to something much bigger than simply choosing the right words, it’s really about creating the kind of environment where the model has what it needs to work properly.

As the technology moves forward, and as people continue to find new ways to apply language models to business, science, and education, this idea of context engineering is likely to become a central part of the conversation. It is not just a different way of saying prompt engineering, it is a more accurate way of describing the careful, layered process of making sure these systems have the right information, in the right form, at the right time.

Read next: Which Industries Rely More On Digital Technologies For The Next Five Years?
by Irfan Ahmad via Digital Information World

Which Industries Rely More On Digital Technologies For The Next Five Years?

There was once a time when businesses across all industries used to run on manual processes, legacy systems and traditional operations. But long gone are those old times with the advent of digital technologies taking shape quicker than ever before in this decade. Industries that were wary of digital technologies are now embracing the new way of operating and hence digital transformation has become the talk of the town, even so with the rapid progress in technologies like Artificial Intelligence and Machine Learning.

In the current market scenario, the question is not about whether or not to implement these digital technologies but how fast these implementations can be done. Even industrial sectors with less digital touchpoints such as the utilities sector are embracing digital technologies like cybersecurity and cloud computing and it's only a matter of time before industries such as agriculture, energy and education start embracing more and more digital technologies. What was once considered futuristic is now part of everyday operations and the speed of innovation has increased so much that keeping up with it has become a task in itself.

This decade has already brought us with so much evolution happening at breakneck speeds and the next five years are set to bring even deeper integration into the operations in companies across all industries. With the vast array of digital technologies like AI, IoT, cloud computing, and automation, companies will shift from experimentation phase to implementation phase by fully embracing change into their business models. In this huge wave of change a question still remains : Which industries rely more on digital technologies for the next five years?

A recent survey published by valantic GmbH asked more than 650 corporate decision makers in the DACH region about the Importance of digital technologies they hold for their company's success in the next five years and not surprisingly AI emerged among the top three technologies in most of the industry sectors surveyed. Only in the utility sector AI was regarded as slightly less important compared to other industrial sectors surveyed. The C-level executives of the utility companies held Green IT in higher ranks compared to other digital technologies. Interestingly, Green IT was in the top only for the utility industry which makes sense since the scope of other digital technologies like AI and Internet of Things (IoT) are seen as a longer term strategy by this industry. The practice of designing and managing information technology in order to reduce its environmental impact is called Green IT. Making IT operations more sustainable while still supporting the business needs is its main goal and hence this ranking by the decision makers suits this industry sector very aptly. Utility companies usually need to meet some environmental standards and hence green IT plays a direct role in helping these companies. In order to make measurable progress towards decarbonization, utility companies are using Green IT which helps them optimize data center efficiency and align with strict sustainability regulations.

On the other hand, Internet of Things or IoT emerged as the top priority across many sectors including Food & Beverage, Retail & Consumer Goods, Healthcare & Pharma and Production Industries. Given that these are mostly physical industries and that they need real time monitoring, traceability and operational efficiency more than anything, these rankings speak for itself. The essential qualities required in such industries would be to maintain quality, compliance and to be competitive and quick. IoT technologies help these businesses track assets, monitor environmental conditions, optimize equipment performance, and respond instantly to disruptions. Whether it's ensuring the freshness of perishable goods, managing inventory levels, improving patient care through connected devices, or reducing downtime on factory floors, IoT offers lots of high-impact benefits. In this competitive market where every minute detail about operational excellence matters, C-level leaders deciding to prioritize IoT for the company's success is the wisest thing to do.

IoT also serves as a foundation to make future innovations such as AI-driven automation and digital twins. Hence this digital technology opens up new possibilities for remote management and smarter decision making which is especially important in the current competitive business environment. Iot becomes more than just a tech trend as companies seek to balance efficiency with agility. The industries that leverage its potential will unlock new levels of innovation that will define leadership in the next decade.

Coming back to the question of which industries rely more on digital technologies, according to the survey, two out of eight physical industries considered for the survey - Food & Beverage and Retail & Consumer Goods are the ones that depend on more digital technologies according to the results. These two industries depend on five digital technologies while the Telecommunications, Automotive, Utilities and Production sectors depend on three digital technologies. While AI is the popular opinion in the retail & consumer goods sector, all five digital technologies in the Food & beverage sector have equal weightage.

Furthermore, sectors like Transportation & Logistics and Healthcare & Pharma rely on four technologies but their preferences vary. Transportation sector holds cloud computing and AI higher than cybersecurity and wireless technologies, while the health sector shows stronger affinity for IoT, which shows the industry's growing use of smart medical devices. Metaverse, blockchain and digital twins, on the other hand, are considered important for the future by relatively few respondents in all sectors. With one exception that also applies to quantum computing. Only in the telecommunications industry is this technology already seen as having great potential by many respondents. The diversity in the survey results show that the priorities of the leaders in the DACH region overall are increasingly driven by digital technologies and certain industries stand out compared to others in higher light but overall progress shows increasing adaptation by businesses regarding digitization.

As we look ahead for the next five years, what is perfectly clear is that digital technologies are no longer confined to early adopters of technology oriented industries. They are indeed the backbone of businesses regardless of sector. From Green IT in utilities and AI driven personalization in retail to IoT powered traceability in food supply chains, the digital revolution has reached all corners of the industrial landscape.

Each industry grows by choosing its own set of digital tools based on their specific operational needs and customer expectations. For example, the utilities sector needs more of Green IT to overcome its challenges and sectors like production and transportation needs more of IoT and cloud to overcome its challenges. This type of growth will eventually lead companies from a state of choosing between digital technologies to a state where they think about how fast these digital technologies can be implemented. In that case leaders will need to balance rapid implementation with responsible management and if they do so will win in the ever growing competitive market.

So, if we are true in understanding how the business world will evolve in this digitally charged decade, we must look at how industries are adapting and implementing different digital technologies. The industries embracing them today are the ones that will lead tomorrow not just in innovation but also in terms of growth and relevance.

Read next: Consumers Are Asking AI Chatbots About More Than Just Tech, New Data Shows

by Irfan Ahmad via Digital Information World