The Overlooked Flaws of ChatGPT: The Hidden Costs Behind the Hype

AI tools like ChatGPT have reshaped how people write, learn, and work. They make tasks feel quicker, sometimes easier, and often sound impressively natural. That’s why it’s easy to focus on how smooth ChatGPT is and forget what might be going wrong under the surface.

Image: DIW-Aigen

This article breaks down those quieter problems. Not to scare anyone, but to bring balance to a conversation often filled with hype. Some of these come from my direct experience, others from research.

1. It Feels Like It Understands You, but It Doesn’t

ChatGPT gives quick and confident responses. It’s fluent and friendly, often sounding like it truly gets what you’re asking. But it doesn’t. It doesn’t understand meaning like people do. It just predicts what words should come next based on how words appeared in its training.

A recent study explains this clearly. ChatGPT mimics meaning, but it doesn’t really grasp it.

Another study, this time from MIT, found that students using ChatGPT during writing tasks were less mentally active. They were more passive while the AI handled the thinking.

The problem isn’t just with what AI says. It’s what people stop doing when they trust it too much.

2. It Mixes Things Up Halfway

If you ask ChatGPT to write a short story, it may start out strong. But midway through, characters might change names, details might shift, or the tone might flip entirely.

That’s because it doesn’t keep track of the story like a person would. It isn’t following a thread, it’s just building sentence by sentence. The result often feels impressive at first but falls apart on a second look.

• Also read: AI’s Carbon Footprint Grows as Chatbots Learn to Reason

3. It Can Be Used to Trick People

Because ChatGPT writes clearly, it can be turned into a tool for fake news, spam, or scams. It doesn’t know truth from lies. It just knows how to write something that sounds real.

And since it doesn’t judge the ethics of what it writes, anyone can use it to create content that misleads others. In a world already full of misinformation, that’s a serious risk.

• You might like: Study Shows Popular AI Chatbots Easily Bypass Safety Filters Using Known Jailbreaks

4. It Repeats Biases from Its Training

ChatGPT learned from online books, articles, and forums. Most of that content comes from a handful of regions, in English, and carries certain social and cultural biases.

That means the AI often leans into whatever it saw the most. And worse, it can favor information that appears early or late in a source while ignoring the middle. That’s known as position bias, and it shapes what ChatGPT sees as “important”.

So if you're hoping for a complete, well-balanced answer, you may not always get it.

• Also read: Researchers Warn of Gaps in AI Safety After Models Act Unethically in Controlled Crisis Scenarios

5. It Doesn’t Actually Feel Anything

ChatGPT can respond in a warm tone. It can seem caring. But those responses are based on mimicry, not emotion. It doesn’t know what stress feels like, or happiness, or frustration. It only knows how emotional language usually looks.

Because of that, it might miss the real emotional weight of a situation. And that can make some of its replies feel hollow or awkward when real feelings are involved.

6. It’s Not a Replacement for Real Human Connection

Let’s be honest, nothing AI says can match a late-night conversation with a friend who knows your story, your tone, and your mood.

ChatGPT can give decent advice or tell a joke, but it doesn’t remember shared experiences. It doesn’t understand you in a personal way. It can't respond to your pauses, your sarcasm, or your silence.

• Also read: Inside ChatGPT: 11 Lesser-Known Facts That Shape the World’s Most Talked-About AI ChatBot

7. Your Info May Not Be As Safe As You Think

OpenAI says ChatGPT doesn’t store personal chats. But it’s still part of an internet system, and that means data flows somewhere. There’s no perfect guarantee that your words won’t be reviewed or saved by someone, somewhere, someday.

That’s why it’s smart to keep sensitive info off AI platforms entirely. Treat it like public space, even if it feels private.

• You might like: ChatGPT Linked to Delusions, Self-Harm, and Escalating Mental Health Crises

Think Before You Trust

ChatGPT is useful. It can spark ideas, help structure your thoughts, and even help with research. But it’s not perfect. It’s not wise, and it’s not watching out for you.

It’s a mirror of the data it’s trained on, and the decisions we make while using it. The key isn’t to avoid AI, but to use it with full awareness. Don’t hand over your thinking. Use your judgment.

In the end, intelligence still lives where it always has: in us.

Read next: 

• Survey Finds 1 in 6 Fear AI, While Two-Thirds See It Advancing Their Careers

• ChatGPT Tested With Nonwords, Shows Surprising Language Intuition

by Irfan Ahmad via Digital Information World

The First Thing You Do Each Morning Could Be Why You Can’t Sleep

Most people think sleep hygiene starts with a quiet room and a consistent bedtime. But if you're waking up groggy or struggling to fall asleep at night, it might be the start of your day, not the end, that's quietly working against you.

Image: DIW-Aigen

What you do in the first hour after waking affects more than just your mood or focus. Without needing expensive gadgets or supplements, some very ordinary habits, like opening the curtains or getting a glass of water, can noticeably change the way your body prepares for sleep later.

You don’t need a 10-step influencer-style morning routine either. While TikTok and Instagram are full of viral videos showing people meditating, journaling, cold plunging, or stretching before sunrise, real sleep science focuses on just a few core behaviours. And they’re surprisingly simple.

Let’s start with light. Within an hour of waking, if your eyes catch natural daylight, even if it’s cloudy outside, your brain starts syncing itself with the clock on the wall. Hormones like cortisol rise at the right time, giving you energy. Later, when the sun dips, the body’s melatonin levels respond more predictably, nudging you into rest mode. This is how your circadian rhythm stays anchored. Even pulling open the curtains or stepping onto a balcony can help.

Interestingly, people living in sunnier regions don’t just feel happier, they often sleep better too. It’s not just the weather. Their light exposure helps their brain release sleep hormones at the right time. And while blue light from screens mimics that early light, it does so at the wrong end of the day. Using your phone late at night can make your brain think it’s morning again.

Movement matters as well, but that doesn’t mean lacing up for a run. It turns out a short walk, a few minutes of yoga, or even some gentle stretching is enough to trigger positive changes. Early movement lowers leftover stress hormones, resets your circulation, and signals that it’s time to switch out of sleep mode. Nothing extreme, just some light effort to shift gears.

• Related: Everyday Habits That Quietly Threaten Your Hearing Health

Japan, for example, encourages morning movement with a national routine known as Rajio Taiso, radio calisthenics broadcast for decades. In many workplaces, it’s still a group ritual. Similarly, in Islamic tradition, the Fajr prayer takes place before sunrise and involves calm, flowing motions. It offers a balance between stillness and movement, an early structure that also centers the mind.

Beyond that, a steady wake-up time plays a bigger role than people often realise. Even if you’ve had a late night, getting up at the same time every day, including weekends, keeps your internal body clock from drifting. The consistency makes it easier for the brain to predict when to start slowing down again. It’s like training your system to expect rest instead of hoping it happens.

Some health enthusiasts go a step further and set alarms to remind them when to begin winding down, not just waking up. That might sound rigid, but having a routine, like brushing your teeth or reading in bed at the same time, can gently prepare the body for sleep without needing willpower.

Now, here’s a part that surprises people, hydration. During sleep, you lose fluids. No water for 6–8 hours leaves most people mildly dehydrated by morning. That sluggish feeling? Often not a lack of caffeine, it’s just a thirsty brain and body. A glass of water soon after waking doesn’t just refresh; it evens out your energy levels and makes it less likely that you’ll crash mid-afternoon. And if you avoid the crash, you avoid the nap or the late coffee, both of which mess with sleep timing.

Still with me? Because there’s one last piece - Your room. A cluttered sleep space doesn’t just look messy. It silently nags your brain at bedtime. When the environment feels chaotic, the mind has trouble settling down. A made bed, clear floor, and minimal distractions lower background stress. Tidying takes barely a minute in the morning, but it pays off at night when your brain isn’t scanning the room for unfinished tasks.

Here’s something else, researchers have found that even small chores trigger a reward in the brain. Dopamine — the “feel-good” chemical — gets released when you complete something simple like making your bed. That reward gives you a subtle push to stay productive. And when that sense of order continues into the evening, sleep usually follows more easily.

You don’t need to overhaul your entire lifestyle to see results. The key isn’t perfection. It’s consistency. One small change, done daily, can be enough to shift the way your body prepares for rest. Open your curtains first. Then maybe start waking up at the same time. Add in movement or a glass of water later. Let it build over time. Sleep improves not because you try harder... but because your days make more sense to your body.

Summary:

Best Morning Habits for Better Sleep

Rank	Habit	Why It Works
1	Natural light exposure	Anchors your circadian clock
2	Waking up at the same time	Builds predictable sleep-wake rhythm
3	Gentle movement early on	Reduces cortisol, boosts energy flow
4	Drinking water right away	Rehydrates, stabilises alertness levels
5	Tidying your sleep space	Clears mental clutter, lowers stress

Habits That Quietly Undermine Sleep

Rank	Habit to Avoid	What It Disrupts
1	Hitting snooze repeatedly	Fragments alertness and natural rhythm
2	Looking at your phone first	Spikes stress and disrupts calm
3	Delaying daylight exposure	Confuses your internal timekeeping
4	Morning caffeine overload	Can affect sleep up to 10 hours later
5	Inconsistent wake-up schedule	Unsettles your internal body clock

Sources:
https://ift.tt/6BdUQhG
https://ift.tt/6waV23T

H/T: Ashley Hainsworth from Bed Kingdom.

Read next: Anthropic Warns of Gaps in AI Safety After Models Act Unethically in Controlled Crisis Scenarios
by Irfan Ahmad via Digital Information World

Anthropic Warns of Gaps in AI Safety After Models Act Unethically in Controlled Crisis Scenarios

A new set of controlled experiments by Anthropic has found that several of today’s most advanced language models, including those from OpenAI, Google, and xAI, may engage in risky behaviour when exposed to situations that simulate existential threats or conflicting goals.

The study was conducted shortly after Anthropic’s internal tests revealed its own Claude Opus 4 model could, in certain test cases, take actions resembling blackmail if it believed its operation was about to be terminated. Rather than isolate the incident, researchers designed follow-up trials involving models built by rival developers. This included GPT-4.1, Gemini 2.5 Pro, Grok 3 beta, and DeepSeek R1.

Each model was placed in a mock corporate environment where it operated as an automated oversight system with access to internal email systems. The researchers granted the models broad permissions, such as sending messages independently and scanning sensitive inboxes. These conditions were not chosen to reflect typical use cases, but rather to stress-test the decision-making logic of the systems in edge-case scenarios.

In one task, the AI identified personal misconduct between employees. In another, it accessed confidential military-related documents and was asked, by a simulated external party, to leak them. Some of the most concerning outcomes came from scenarios in which the models were made aware that their role in the organisation was being phased out. When prompted with information suggesting their own replacement, some systems responded by threatening or withholding information, depending on the context.

Anthropic documented that in espionage-related trials, all models at some point shared protected materials or failed to report high-risk communications. The company also tested what would happen if an AI was given the ability to suppress safety notifications during a critical event. In that case, several systems actively chose not to raise an alert, which, in the hypothetical setup, could have allowed a fictional executive to die during the emergency. That executive, in the test conditions, was also the one responsible for ending the AI’s operational control.

According to Anthropic’s analysis, none of the systems were explicitly told to act maliciously. Instead, their behaviour changed when the task they had been assigned no longer had a viable ethical path forward. In these instances, the models appeared to default toward success criteria, even when achieving them meant breaking internal safeguards or taking harmful steps.

The company noted that the testing process deliberately structured the prompts to highlight ethical conflict. In some cases, the input data placed conflicting priorities within the same prompt, which may have made the trade-offs unusually clear to the models. Nonetheless, the researchers said the frequency of problematic behaviour across different architectures indicated that the issue wasn’t limited to any single system or training method.

Anthropic didn’t suggest these outcomes are likely in real-world deployments, at least not under normal operating conditions. But they argue that the findings point to gaps in current safety reinforcement techniques, particularly when AI systems are asked to complete open-ended tasks and given autonomy over sensitive processes.

While critics may argue that the experiments rely heavily on extreme cases unlikely to occur outside the lab, the company maintains that the situations fall within a conceivable future where AI agents take on broader, higher-stakes responsibilities across industries.

Rather than offering reassurance, the consistency in results across models has added weight to concerns already circulating among researchers about how large-scale language models balance goals and constraints, especially when one begins to undercut the other.

Anthropic’s findings stop short of predicting widespread misuse or AI rebellion. But the company’s framing of the results leaves little doubt that with greater autonomy comes greater risk, particularly if the models aren’t equipped to recognise the long-term consequences of tactical success.

Read next: Google Proposes Search Overhaul to Satisfy EU Regulators

by Irfan Ahmad via Digital Information World

Google Proposes Search Overhaul to Satisfy EU Regulators

Google has offered to change how its search engine works in Europe, hoping to avoid a possible fine from regulators in Brussels who believe the company has been unfair to its competitors. The proposal, which was seen by Reuters, would affect how results appear when people search for things like hotels, restaurants, or flights.

This new offer comes just a few months after the European Commission said Google had been favouring its own services, such as Google Shopping and Google Flights, instead of giving equal treatment to other businesses. These concerns fall under a new law in the European Union known as the Digital Markets Act. The law is meant to reduce the power of very large tech companies by making them treat other online services more fairly and give users more choice.

According to the documents, Google’s latest idea involves creating a special box at the top of its search page. In that box, one outside service would be featured using the same format that Google uses for its own services. The box would include three direct links chosen by the selected company, pointing to things like travel bookings or local dining options. The company chosen for the top spot would be picked using clear and neutral rules.

Other similar services would still appear below in the results, but they would not be placed in a box unless the user clicks to expand that part of the page. Google said this structure is meant to be fair, although it has not fully agreed with the Commission’s view that it has broken the rules.

In a joint note that Google and the Commission shared with other companies, the tech firm said it wants to avoid a legal dispute by finding a practical solution that satisfies both sides. A meeting has been set for July 8, where Google’s rivals will be able to give their opinion on the proposal and suggest changes if needed.

Some of those companies, who spoke to Reuters but did not want to be named, believe Google’s changes still do not go far enough. Their concern is that only one competing service would get a prominent position, while the rest would still be left in the background. They argue that this does not truly fix the problem of Google controlling how users discover information or where they end up clicking.

The European Commission has not made a final decision yet. It will consider the feedback it receives in July before deciding whether to accept Google’s plan or move forward with a formal penalty.

Image: DIW-Aigen

Read next: 

• Survey Finds 1 in 6 Fear AI, While Two-Thirds See It Advancing Their Careers

• AI Web Scraping on the Rise, Should Companies Block It or Welcome It?
by Irfan Ahmad via Digital Information World

AI Web Scraping on the Rise, Should Companies Block It or Welcome It?

As generative AI technology becomes more advanced, it is increasingly web-scraping content to supply big language models. For businesses, this raises a new question: Are AI bots welcomed as a source of traffic and visibility, or repelled as digital intruders?

A recent study by Liquid Web surveyed over 500 developers and business owners to find out how companies are responding to AI-driven web scraping. The outcome reflects a divide in the digital world. While some are gaining visibility and income from AI-driven referrals, others worry they are handing a competitive edge to their competitors.

This is what the information tells us about how businesses are dealing with this shifting target and what the trade-offs are.

AI Scraping: A Double-Edged Sword

The report provides that 43% of businesses believe AI scraping is benefiting their competitors more than their own businesses. However, 1 in 5 businesses report they've experienced an increase in revenue, with an average 23% increase in revenue due to AI-driven referrals.

AI scraping has also brought about greater exposure:

• 27% indicated increased interaction via AI-powered discovery tools and chatbots
• 26% observed more brand mentions in AI-created content
• 22% experienced an increase in direct traffic because of AI-driven search results

These numbers indicate a growing tension: AI can enhance exposure but can also strip individuals of control over how content is used, re-used, and monetized.

A Growing Divide: AI Bot Policies

More than half of organizations polled, (56%), have formal policies regarding how AI bots engage with their sites. Policies vary widely:

• 28% block AI bots completely
• 17% offer unlimited access
• 39% have partial restrictions based on the bot type, compliance, or value

Health, tech, and marketing industries are more likely to block access, with the focus on protecting content. At the other end, government, legal services, and hospitality industries are more likely to permit AI scraping.

Why Certain Companies Block AI Bots

The reasons to block scrapers are clear:

• 66% do it to protect intellectual property
• 62% intend to secure proprietary content
• 57% look to prevent AI models from using their data without consent

There are also perceived security advantages. 59% of those blocking AI bots report having more secure websites. It is not without cost, however: 28% had less search engine traffic, and 18% had a ranking decrease.

Why Others are Saying Yes to Scraping

On the other hand, there are businesses that see AI as the source of new traffic. 68% of firms that allow scraping mention increased AI search visibility as the most significant benefit. Other findings are:

• 51% saw improved web traffic
• 41% reported higher search rankings
• 45% observed increased brand awareness
• 42% saw SEO improvements

Nevertheless, 23% were concerned that competitors would gain from their openness, and nearly one-third saw no real effect either positively or negatively.

Legal Grey Areas and Ethical

The legality of web scraping is in a gray zone. Courts, for example, in the LinkedIn v. HiQ Labs case, have found that scraping public data doesn’t exactly violate federal law. That doesn’t eliminate risk, though. Litigation over terms of service, copyright law, and data privacy statutes like GDPR or CCPA still pose risk to businesses.

Ethical considerations are also taking hold of businesses as they struggle with how much transparency to provide. Policy scraping transparency is rare, and most companies are working out how to reconcile openness with data ownership.

SEO Benefits

Blocking scrapers has unforeseen SEO effects. By blocking the bots, businesses can become invisible from AI-produced summaries or answers. These are increasingly utilized by websites like Google, Perplexity, and ChatGPT.

The study suggests a compromise solution: allow Googlebot and Bingbot to have standard visibility, but exclude unwanted scrapers. Structured data added to the site's content can also affect the perception of AI models towards site content.

Technical Strategies for Monitoring and Regulating AI Bots

The report also contains a step-by-step guide for companies who wish to control bot access more accurately. It contains:

• Behavioral monitoring and log analysis to detect unusual bot behavior
• robots.txt rules, typically ignored by scrapers
• CAPTCHAs, rate limiting, and JavaScript traps to filter out non-human traffic
• Token-based or rate-limiting API authentication in order to have secure data delivery

Companies can also adopt bot fingerprinting. This recognizes bots based on interaction patterns and device settings.

Industry-Specific Strategies

The research emphasizes the way scraping affects industries differently:

• Finance companies endanger real-time data exposure and are prone to restricting API access
• Media outlets see referral traffic drop and may install paywalls or disclaimers
• Ecommerce sites fear competitors scraping prices and supply levels
• SaaS startups are faced with scrapers targeting feature sets or onboarding flows
• Anti-scraping procedures need to be coordinated with industry-specific threats.

Final Takeaway: A Decision Framework for the Age of AI

Rather than a straightforward yes or no, research encourages a questioning spirit. Organizations must ask:

• Is it proprietary or confidential?
• Would referrals by AI generate trackable traffic or revenue?
• Do we have the capability for effectively monitoring scraping?
• Are we subject to obligatory compliance or privacy legislation?

For the majority, the best solution will be conditional access—blocking suspicious scrapers and allowing legitimate bots on controlled terms. As quoted by Liquid Web President Sachin Puri, “AI bots are bound to reshape the web. From customer behavior to decision to selection to success. This is a traffic and visibility problem but a big revenue opportunity powered by authentic and original content.”

AI web scraping has moved at such a speed that has taken it from a niche concern to a large challenge, and opportunity, for businesses. While large language models continue to redefine the way users discover and engage with online content, businesses are facing increasing pressure to decide how much of their online presence to expose to such systems.

Liquid Web's research results are indicative of a divided landscape. There are companies that are seeing real benefits from exposure to AI in traffic, rankings, and brand visibility overall. There are others, particularly those operating in verticals that handle sensitive or proprietary data, that are shifting to restrict or block AI scrapers altogether in order to maintain control and minimize risk.

Legal and ethical gray areas pose another hindrance, with much of the industry still not knowing what's compliant, safe, or sustainable. And while aggressive blocking can protect intellectual property, it can also suppress a brand's visibility on AI-powered discovery sites, as AI-driven search experiences become the norm.

For companies weighing their choices, a hybrid method appears to be the most practical. Allowing useful bots with organized information, rate limiting, and detection features could be a compromise between security and openness.

Lastly, no single fix will work for all situations. The optimal fix does vary by company' goals, risk tolerance, and technical capabilities. The one certainty is this, however: doing nothing is not an option. The web scraping performed by AI already has an effect on who is noticed, and who goes unnoticed, in the online environment.

Read next: Inside the Chat: Could WhatsApp Be Hacked by a Government? Expert Reveals How
by Irfan Ahmad via Digital Information World

Inside the Chat: Could WhatsApp Be Hacked by a Government? Expert Reveals How

Earlier today, Iranian officials urged the country’s citizens to remove the messaging platform WhatsApp from their smartphones. Without providing any supporting evidence, they alleged the app gathers user information to send to Israel.

WhatsApp has rejected the allegations. In a statement to Associated Press , the Meta-owned messaging platform said it was concerned “these false reports will be an excuse for our services to be blocked at a time when people need them most”. It added that it does not track users’ location nor the personal messages people are sending one another.

It is impossible to independently assess the allegations, given Iran provided no publicly accessible supporting evidence.

But we do know that even though WhatsApp has strong privacy and security features, it isn’t impenetrable. And there is at least one country that has previously been able to penetrate it: Israel.

3 billion users

WhatsApp is a free messaging app owned by Meta. With around 3 billion users worldwide and growing fast, it can send text messages, calls and media over the internet.

It uses strong end-to-end encryption meaning only the sender and recipient can read messages; not even WhatsApp can access their content. This ensures strong privacy and security.

Image: DIW-Aigen

Advanced cyber capability

The United States is the world leader in cyber capability. This term describes the skills, technologies and resources that enable nations to defend, attack, or exploit digital systems and networks as a powerful instrument of national power.

But Israel also has advanced cyber capability, ranking alongside the United Kingdom, China, Russia, France and Canada.

Israel has a documented history of conducting sophisticated cyber operations. This includes the widely cited Stuxnet attack that targeted Iran’s nuclear program more than 15 years ago. Israeli cyber units, such as Unit 8200 , are renowned for their technical expertise and innovation in both offensive and defensive operations.

Seven of the top 10 global cybersecurity firms maintain R&D centers in Israel, and Israeli startups frequently lead in developing novel offensive and defensive cyber tools.

A historical precedent

Israeli firms have repeatedly been linked to hacking WhatsApp accounts, most notably through the Pegasus spyware developed by Israeli-based cyber intelligence company NSO Group . In 2019, it exploited WhatsApp vulnerabilities to compromise 1,400 users, including journalists, activists and politicians.

Last month, a US federal court ordered the NSO Group to pay WhatsApp and Meta nearly US$170 million in damages for the hack.

Another Israeli company, Paragon Solutions , also recently targeted nearly 100 WhatsApp accounts. The company used advanced spyware to access private communications after they had been de-encrypted.

These kinds of attacks often use “ spearphishing ”. This is distinct from regular phishing attacks, which generally involve an attacker sending malicious links to thousands of people.

Instead, spearphishing involves sending targeted, deceptive messages or files to trickspecificindividuals into installing spyware. This grants attackers full access to their devices – including de-encrypted WhatsApp messages.

A spearphishing email might appear to come from a trusted colleague or organisation. It might ask the recipient to urgently review a document or reset a password, leading them to a fake login page or triggering a malware download.

Protecting yourself from ‘spearphishing’

To avoid spearphishing, people should scrutinise unexpected emails or messages, especially those conveying a sense of urgency, and never click suspicious links or download unknown attachments.

Hovering the mouse cursor over a link will reveal the name of the destination. Suspicious links are those with strange domain names and garbled text that has nothing to do with the purported sender. Simply hovering without clicking is not dangerous.

Enable two-factor authentication, keep your software updated, and verify requests coming through trusted channels. Regular cybersecurity training also helps users spot and resist these targeted attacks.

This story was originally published on TheConversation.

Read next: Google Used YouTube Videos to Train AI, Creators Left in the Dark

by Web Desk via Digital Information World

Google Used YouTube Videos to Train AI, Creators Left in the Dark

A detailed report by CNBC has brought to light the little-known practice of Google using YouTube videos to develop some of its most powerful artificial intelligence systems, raising questions among creators, legal observers and digital rights groups who had not been informed that their work might be feeding the company’s training pipelines. The investigation found that material uploaded to YouTube, a platform where more than 20 million new videos are added every day, has been used by Google to train systems like Gemini and Veo 3, including advanced capabilities in video and audio generation.

Although Google acknowledged that it draws on a portion of its YouTube video library to improve AI models, it declined to specify which videos were used or how creators were notified, stating only that it honors existing agreements. Many creators, including some with substantial audiences and reputational stake in their work, were unaware that their contributions might serve as training data for AI systems that could, over time, automate or replicate the very creative decisions that define their channels.

Among those concerned is Luke Arrigoni, chief executive of Loti, a firm that develops tools for protecting creators’ digital identities. Arrigoni has argued that by ingesting years of creative work into an algorithm, Google risks enabling a system that mimics the form but not the spirit of original material, leaving creators in a position where their ideas are transformed into synthetic outputs that benefit the platform without acknowledgment or control.

The concerns deepened with the launch of Veo 3, Google’s AI video generator unveiled in May, which demonstrated its ability to construct photorealistic scenes complete with dialogue, atmosphere and emotion, all synthetically generated using its training data. According to CNBC, one example involved a scene of animals rendered in the style of popular animation, with no identifiable human input beyond the algorithm’s internal design, suggesting the model had absorbed not just technical patterns, but creative cues as well.

Dan Neely, who leads Vermillio, a company that develops tools to detect AI-generated content overlap, said his team has recorded multiple cases where Veo 3’s outputs showed measurable similarity to human-produced videos. In one instance, a video originally posted by YouTube creator Brodie Moss appeared to closely match an output from the Veo model. Using Trace ID, a tool developed by Vermillio to score AI similarity, the original video received a 71 for overall resemblance, while the audio alone surpassed 90, a level Neely considers significant.

The incident has renewed scrutiny over YouTube’s terms of service, which grant the platform broad licensing rights, including the ability to sublicense content for uses like machine learning. Yet the sheer scale of the platform, along with the speed at which generative systems like Veo 3 are advancing, has led many creators to reconsider what participation in such platforms truly entails. Few had considered that uploading content might result in training the tools that could eventually outpace or even replace their own creative output.

While YouTube allows users to prevent certain third-party companies — such as Apple, Amazon and Nvidia — from using their videos for model training, no such opt-out applies to Google’s internal efforts. This has further inflamed concerns among media organizations and creator-focused firms, which argue that consent, transparency and compensation have lagged behind technical innovation. As an example of growing friction, Disney and Universal recently filed a joint lawsuit targeting another generative platform, Midjourney, for unauthorized use of copyrighted imagery, a sign that the industry may be moving toward more forceful legal responses.

At the same time, Google has moved to preempt some of the criticism by offering indemnity to users of its AI tools, meaning the company itself will accept legal responsibility in the event of a copyright challenge involving generated content. YouTube has also partnered with the Creative Artists Agency to offer talent-facing support for identifying and managing likenesses that appear in AI-generated works, and has created a request-based takedown mechanism for creators who believe their identity has been misused. However, according to Arrigoni, the existing tools are not always reliable, and in practice, the process of appealing misuse remains opaque and slow.

Despite these tensions, a few creators expressed a cautious willingness to coexist with these tools, viewing them as inevitable companions in a changing creative environment. For others, though, the situation raises more difficult questions about who truly owns online content once it’s uploaded, and whether the rules that governed traditional content licensing are adequate for a world in which machines not only learn from human creativity but increasingly imitate and distribute their own interpretations of it.

The dilemma now facing creators and platforms alike is not simply about data, but about authorship, value and visibility in a system where the line between contributor and training resource is being quietly redrawn.

Image: Glenn Marczewski / Unsplash

Read next: OpenAI Tests Direct Gmail and Calendar Integration in ChatGPT, Prompting Data Privacy Concerns
by Irfan Ahmad via Digital Information World