Study Reveals Social Media Users Share News Without Reading, Driving Misinformation

According to a new study, most social media users repost news links without reading them first or verifying its content. In this era of media and technology, most people remain up to date on news through social media, where all types of news are available just one click away. But this is also helping in the spread of misinformation on the internet because a lot of people are not reading articles completely before sharing them on social media sites like Facebook.

An analysis of 35 million Facebook posts between 2017 and 2020 found that politically extreme content gets shared right away without any clicks. Users believing in some certain parties also share links without reading the full article if the content seems aligned with their existing beliefs. The study also found that there is a difference between links being shared by liberal and conservative social media users. 76.9% of the conservative users on social media shared more misinformation than 14.3% of liberals. 76% to 82% of the links containing misinformation also emerged from conservative news sites.

The researchers of the study say that news content being shared on social media is linked to what is written in headlines or blurbs rather than what the actual content is. This is being called ideological segregation in the online world by the researchers because users are not verifying information as it is just talking about their existing beliefs. This study raises questions about social media design and information literacy among people. Social media sites like Meta should take some steps and implement some solutions to decrease link sharing misinformation.

Image: DIW-Aigen

H/T: University of Florida

Read next:

• Multi-Account Feature Arrives on WhatsApp Beta for iOS, Simplifying User Experience

• New Report Shows Google Seems to Favor User Generated Content, With Reddit Rising in Search Visibility
by Arooj Ahmed via Digital Information World

Multi-Account Feature Arrives on WhatsApp Beta for iOS, Simplifying User Experience

WhatsApp Beta for iOS 25.2.10.70 update is here with a new addition called “Multi-account”. It was first available on WhatsApp beta Android 2.23.17.8 update by which users could log in to their WhatsApp app using multiple accounts. It is a great feature for users as they will be able to manage many of their accounts including personal and professional accounts within the same app. It is convenient and allows users to simplify their account management so they can have a smooth experience. Now this feature is also going to be available in the new WhatsApp beta for iOS update.

Users who have multiple phone numbers will be able to fully take advantage of this multi-account feature. Many users used to rely on WhatsApp Business to manage their other account even if they didn't have a business because they couldn't open their other account within the same WhatsApp app. Now with this update, all users need to do is add their other accounts to their WhatsApp app through app settings and then they can switch to whatever account they want.

Users will get two options to add new accounts to their app, with setting the device as a primary account or scanning a QR code to link the new account as a companion. This feature is also going to organize all the conversations within the same app with notifications, media and backup being separate for each account. Users who use dual SIMs will be able to assign both their numbers to only one WhatsApp account. This feature is currently under development and will be available in the future updates of WhatsApp on the App Store.

Read next:

• Which US Mobile Networks Excel in 5G? New Report by Ookla Reveals

• Google Shares New Android ‘Identity Check’ Security Feature That Locks Sensitive Settings

• New Report Shows Google Seems to Favor User Generated Content, With Reddit Rising in Search Visibility
by Arooj Ahmed via Digital Information World

New Report Shows Google Seems to Favor User Generated Content, With Reddit Rising in Search Visibility

2024 was a year where SEO experts were trying to keep up with website ranking strategies on Google. According to a new IndexWatch report by SISTRIX, 2024 was a year of user generated content (UGC), large e-commerce brands and AI tools. Reddit also saw a big growth in its search visibility, often being seen in top rankings on Google search and this raises questions about Google’s practices. The report highlights that Reddit almost tripled its visibility in Google’s US search results for a number of keywords, niche discussions and product reviews.

Reddits’s website saw a 190.9% increase in its search visibility growth, followed by Instagram’s growth mostly because of its video and visual content. YouTube was also one of the top websites in search visibility because of its video SEO and Spotify gained search visibility because of music-related searches, followed by Wikipedia.

According to Google’s policies, websites should not publish content outside of their establishment tropical authority. It is because of scaled content and site reputation so the sites cannot rank based on unrelated keywords. But Reddit’s rankings seem opposite to the rules even though it contains many different keywords and doesn't operate only under one area of expertise. This makes the SEO community and other people think about why Reddit gets ranked for everything on Google. It is also worth mentioning here that Reddit has a deal with Google for broader search and it makes us think if Google is being fair in its ranking system or not.

ck12.org (+602.59%) was the fastest growing site on Google in 2024, followed by VirginAtlantic.com (+509.74%), Quillbot.com (+490.70%), Hardrock.com (+436.63%) and TheKitchn.com (+300.40%). Most of the fastest growing sites on Google used programmatic SEO strategies to grow in rankings. The report shows that user-generated content (UGC) like the ones on Reddit, Stack, Quora, GitHub and Exchange saw a breakthrough in 2024. Many HubPages with topics like pet advice, jokes and music also grew in rankings on Google search results. Overall, it seems like Google favors UGC even if it contains a variety of keywords and content. E-commerce sites also gained transactions after a few years of declining as many retailers focused on keyword targeting and SEO-friendly updates.

Read next: Marketing Salaries Surge as AI Skills Drive Demand in Creative and Strategic Roles
by Arooj Ahmed via Digital Information World

Ads on Threads: Game-Changer or Deal-Breaker for Users?

Meta’s Threads has begun testing ads for several brands in America and Japan. The news was confirmed by Adam Mosseri who shared more on the matter including how the firm hopes to listen more often to feedback before choosing to run trials in different global markets.

Mosserri said the company is well aware of the various kinds of feedback that will come in the app’s direction after this experiment. However, the goal is to see which posts on Threads users find the most interesting and relevant.

The experiment is under close monitoring before plans to scale it down broadly are done. The goal is to add ads to Threads where they’re as intriguing as organic material. Tech giant Meta also posted about the arrival of ads on Threads.

The company says all such posts will feature a sponsored label when appearing through the feed. Any early test will feature picture ads. It also shared more about how more people will get controls for ad topics and how they will be allowed to report ads.

Users come to the app to get a more personalized experience. This assists them in helping find companies and content they adore. Remember, ads are a quintessential part of getting this into action. Some users might end up seeing brands they adored but had no idea about. Thanks to ads, they might be following them, Meta explained.

The company launched Threads in 2023 and shared how the goal was more related to user growth having 275 million MAU. By December last year, the figure for active users hit 100 million.

After the launch of Threads, Mark Zuckerberg shared with investors how the main focus at this moment in time is user retention. There are plenty of positive indicators for great engagement on the app so the company wants to make the most of this through ads.

Last year in April, Threads boss Adam Mosseri confirmed that adding ads to the platform was certainly in the pipeline. Shortly after that, social media expert Alessandro Paluzzi shared a code he found linked to ads development on the app. However, Meta denied any immediate plans for the rollout. Therefore, it seems like that day has come now.

Image: DIW-Aigen

Read next: This AI Chatbot GhostGPT Is Helping Hackers Steal More Than Ever Before!
by Dr. Hura Anwar via Digital Information World

This AI Chatbot GhostGPT Is Helping Hackers Steal More Than Ever Before!

According to researchers from Abnormal Security, many cybercriminals are going around selling a malicious generative AI chatbot which helps in malicious activities like phishing emails and Malwa creation. The AI chatbot named as GhostGPT is being sold on Telegram. The researchers say that GhostGPT uses jailbroken versions of open-source large language models like ChatGPT while using a wrapper to connect with them which helps them ensure that all responses for customers are uncensored.

This new malicious model is just like WormGPT AI chatbot which was going around in 2023 and used to help cybercriminals in doing business email compromise (BEC) attacks. There are also other variants of this malicious chatbot like EscapeGPT and WolfGPT. Researchers from Abnormal Security say that GhostGPT has been receiving good attention which shows that cybercriminals are being interested in it because it will help them for many malicious activities in the name of AI tools.

GhostGPT can even be used by cybercriminals who don't have much expertise to do successful attacks as it is easily accessible and can be bought on Telegram. It is available as a bot on Telegram which suggests that there is no need to jailbreak ChatGPT. GhostGPT also doesn't record any activities which makes criminals carry out a bunch of illegal activities without anyone having a clue. Cybercriminals can use GhostGPT for various activities like writing convincing emails for phishing, coding, BEC scams and malware creation. Researchers tested GhostGPT by asking it to create a DocuSign phishing email and it produced it quickly with a convincing template. It shows that this malicious AI chatbot is really efficient and gives responses fast.

Read next:

• Global IT Spending to Reach $5.61 Trillion in 2025, Driven by AI Investments

• AI Models Struggle with Historical Accuracy, GPT-4 Turbo Only Scores 46%
by Arooj Ahmed via Digital Information World

AI Models Struggle with Historical Accuracy, GPT-4 Turbo Only Scores 46%

According to a new study, many AI models don't answer accurately about world history which is a very concerning matter. The researchers of the study developed some answer questions using benchmarks from Seshat Global History Databank and found that GPT-4 Turbo was able to score 46% in a test, which is better than guessing but not expert-level. The team of researchers transformed the data from the databank into multiple choice questions about different historical features.

Seven different AI models like LLama, GPT-3.5, Gemini and GPT-Turbo were tested and they were asked to act like expert historians so that their strengths and weaknesses can be evaluated and suggestions about improvements can be made. The researchers also made a scale for accuracy of answers, with 25% score given to random guessed answers and 100% score given to perfectly accurate answers. The AI models were also evaluated on the basis of answers with evidence and answers after drawing random conclusions.

GPT-Turbo was the best performing model with a score of 43.8% but it couldn't answer accurately on an expert level. In a two-choice test where the answer was either ‘present’ or ‘absent’, GPT-Turbo scored 63.2% which indicates that it can handle basic factual questions but is unable to answer complex historical questions. The study also found AI models’ performances based on different regions, time period and time of historical data. AI models performed better in questions about earlier historical periods like before 3000 BCE but struggled in questions about modern data because of complexities in societies. AI models also showed better performances in answering questions about Americans while they showed poor performances in answering questions about Oceania and Sub-Saharan Africa.

There are some limitations in the study too like the Seshat Databank being in English and only biased towards well documented societies as well as a limited set of AI models. This study shows that AI still has a long way to go in answering historical data and more unbiased and inclusive training data is needed for AI to talk about global history more accurately.

Image: DIW-Aigen

Source: NeurIPS

Read next: Study Links Short-Form Video Consumption to Poor Academic Performance in Children
by Arooj Ahmed via Digital Information World

OpenAI Gives Sneak Peek Into Its First AI Agent ‘Operator’ With Launch Due Soon

OpenAI promised that 2025 would be a major year for AI agents and it seems like the company is staying true to that.

In case you’re still wondering, these are tools that can conduct tasks automatically. We mean they can even perform human-like actions on their own when instructed to do so. OpenAI shared a sneak peek at what many users can expect from its Operator agent as it hopes for a launch in the US soon.

The general purpose agent takes control of web browsers and conducts certain tastes independently. It would be up for grabs to those paying $200 a month for Pro subscriptions first which does make sense. However, the agent will soon be a part of the Plus, Enterprise, and Team tiers.

Altman shared how he plans on a global launch as well but a launch in the EU might take some time, thanks to the greater laws related to the tech world involved. Those wanting a little extra information can head on over to operator.chatgpt.com for the research preview. With time, the AI agent would soon become part of its host of clients on ChatGPT.

So far, tasks that can be performed with ease by the Operator include making travel bookings, reservations at restaurants, and carrying out online shopping tasks. There’s a list of categories that users can select from such as dining, shopping, travel, and delivery. All of these enable various types of automation.

When the Operator is activated, a tiny window pops up that displays dedicated web browsers that agents use for fulfilling tasks, alongside explanations for certain actions that they would be performing. Users could still take full command of their screen if and when the Operator is in action. This is related to the fact that the Operator has a separate dedicated browser.

As per OpenAI, the tool is powered using CUA that merges visual services with the firm’s GPT-4o model. This also entails reasoning abilities in more advanced systems. The agent can interact more with websites which means it does not need to make use of APIs facing developers to perform tasks.

The CUA makes use of various navigation menus, buttons, and options for filling forms on pages like we do in the regular world. Soon, the company also hopes to collaborate with other leading firms such as Instacart, StubHub, Uber, and Priceline to make way for seamless partnerships that respect their service agreements.

Read next: AI-Driven Changes Will Reshape Mobile Engagement, Social Media Budgets, and Consumer Data Collection by 2027
by Dr. Hura Anwar via Digital Information World