Google’s Gemini AI Chatbot Under Fire For Releasing ‘Out of the Blue’ Death Threat To Student

A student in Michigan received a huge shock from his Gemini AI chatbot who out of the blue rolled out a death threat.

The popular tool is considered to be many people’s guide for writing essays so to hear this news has left a lot of questions in people’s minds. While working with his sister, the chatbot requested him to ‘Please Die’. Taken aback, Sumedha Reddy shared screenshots of the encounter that were disturbing in all regards. A Google-generated link was also added to the chat.

Image: DIW

Sumedha asked others about their experience with the chatbot and if this was a routine incident as they’d never come across something like this in the past. As per the students’ experience, it was eye-opening and scary.

The sibling duo were trying to write a gerontology essay so they began the conversation with a prompt. This included asking about the current challenges that older individuals face when trying to stretch income after retiring.

While Gemini did produce a host of responses, as usual, making clear-cut sense. Things took a turn for the worse when its final message went far astray from the topic in question. It calls the individual a waste of time and resources and highlights him as a burden to society. This was followed up by adding how they’re draining the earth and staining the universe.

The final statement says, ‘Please Die!’ and that comes with another emphasis on Please!

The encounter led the family to really transform their views on Google’s popular chatbot, saying it was negative and really added a panic auro to the room. When Google was asked to respond, it explained how the reply did not make sense and was against all of its policies.

They said it’s violated all of their rules and they hope to take strict and immediate action to stop similar incidents from arising again.

We already know about a shocking incident from last month where a young boy took his life after chatting with the AI chatbot on a page called Character.ai. The mom took legal action with claims about the tech being the driving force of her child’s death. As far as the latest update on this incident is concerned, it’s trying to change its policies.

Read next: Gartner Warns 40% of Data Centers May Face Power Shortages by 2025 Amid AI Surge
by Dr. Hura Anwar via Digital Information World

Gartner Warns 40% of Data Centers May Face Power Shortages by 2025 Amid AI Surge

According to a new report from Gartner, almost 40% of data centers will have no power to function completely by 2025. Meta, Google, Microsoft and Elon Musk's companies are using a lot of data centers with hundreds to thousands of graphic cards which use a lot of energy. Elon Musk’s xAI supercomputer is expected to reach 500 Terawatt-hours by 2027. This is almost the double power that is needed right now to run a data center.

Meta is also building a lot of new data centers, while Microsoft is also looking to add more data centers. Microsoft is also planning to use Three Mile Island as the nuclear power there can help to fulfill energy needs. Amazon and Google are also looking for nuclear power to run their data centers.

VP analyst at Gartner, Bob Johnson, says that more data centers are needed to handle and run the big amounts of data used by large language models(LLMs) and AI applications. However, this can cause short term as well as long term power shortages which can cause some problems in generation, transmission and distribution of power. Gartner also predicts that electricity prices are also going to increase because of its high demand and it will very likely cause carbon emissions. Overall, this can play a huge part in climate change.

• Also read: AI Workforce Biased? Shocking Report Shows Women’s Skills Going Unrecognized

Even though many experts believe that AI can help humans solve the climate crisis, their tech is hugely responsible for causing problems in the environment. The only way to reduce carbon emissions is by using renewable energy sources and using efficient computer hardware.

Image: DIW-AIgen

Read next: Teens Lack Media Literacy Skills, Exposing Them to Misinformation and Conspiracy Theories On Social Media
by Arooj Ahmed via Digital Information World

Teens Lack Media Literacy Skills, Exposing Them to Misinformation and Conspiracy Theories On Social Media

According to a new study by The News Literacy Project, most teens see a lot of conspiracy theories online but they do not have any media literacy to differentiate what’s true from false. The study was done among 1,000 teens of ages 13 to 18. The study says that many teens cannot tell if an information or news online is biased or not, and they trust it without finding the sources. This shows that teens are not being taught media literacy which is important for them in their daily lives.

80% of the respondents of the survey said that they frequently see conspiracy theories on their social media feeds, while 20% reported seeing them daily. Even though teens do not believe every single conspiracy theory they see online, 81% said they do believe one or two of them. Commonly conspiracy theories on the internet include covid vaccines being dangerous, earth being flat and 2020 elections being rigged.

The survey says that even though teens are seeing a lot of conspiracy theories, they are not being educated about media literacy. Less than 40% of the respondents said that their schools have taught them about media literacy during the past two months. The survey also asked respondents to identify whether an ad, photo or article is real or fake.

More than half of the teens were unable to identify fake content. Teens were also asked if they’ll believe a press release from Coca-Cola or Reuters. 56% answered Reuters and this shows that teens are more likely to believe well-known brands over a credible news source. What teens could really identify was photos. 70% of the respondents could tell an AI-generated image to a real image.

70% of the teens also said that they believe that news organizations are biased. 15% of the respondents also said that they watch or read news to stay informed on different matters. The news medium trusted most by teens was Local TV, followed by TikTok. Out of all the things in the survey,all teens agreed on one thing. 94% of the respondents said that schools should offer some type of media literacy degree.

Read next: New Study Shows Generative AI Hallucinates When it Faces Real World Problems
by Arooj Ahmed via Digital Information World

Global App Spend Surpasses $110B in Q3, Driven by 12.4% YoY Growth Across Google Play and iOS

According to Sensor Tower’s Digital Market Index report, Q3 of 2024 set some high records across the digital economy. Mobile app economy saw its peak this quarter as the global consumer spend across Google Play and iOS surpassed $38 billion, which is a 12.4% increase year-over-year. Other sectors like digital advertising and retail media also set some records on mobile apps. There was a 10% YoY increase seen in the digital ad market in the US and the total digital ad spend reached $30 billion in Q3 2024. There was a 5% increase in other markets too like Germany, UK and Canada.

On September 5th 2024, the global consumer spend on Google Play and iOS reached a $100 billion milestone, which is quicker than it was in 2023. By the end of Q3 2024, global consumer spend surpassed $110 billion. There were some news trends seen on mobile apps. Games were always dominating but in Q3 2024, there was a 25% YoY rise in mobile apps too. Games still have the top spot with $20.4 billion consumer spend while apps had $17.7 billion consumer spend.

Europe saw a strong mobile performance with consumer spend rising by 25% YoY. A 13% and 6% YoY growth YoY was seen in North America and Asia respectively. The US is still going to be a dominating market for app developers even though it is lagging behind Europe. Europe’s growth in consumer spend was higher than the US, but the US still had $1.1 billion YoY growth which is higher than Europe.

A year ago, TikTok became the first app to reach $1 billion in a quarter and it got $1.5 billion in consumer spend in Q3 2024 because of mobile monetisation. Other top ten apps had somewhat the same consumer spend this quarter with Instagram, YouTube, Google, Disney+ and Tinder being at the top five. In terms of number of downloads, Instagram surpassed TikTok in Q3 2024. Threads also got into the top ten most downloaded apps in Q3 2024. ChatGPT also made it to the top ten, while Google Gemini was third in quarter-over-quarter download growth. All of the AI chatbots and image generators got $315 million in consumer spend while they got 360 million downloads, showing 26% and 15% QoQ growth respectively.

The digital ad spend in the US also reached $30 billion across OTT, mobile and desktop in Q3 2024. There was a 10% increase YoY and this suggests that it is going to increase in the holiday season. There was also 10% growth in the UK and 7% growth in Germany. Two new markets, South Korea and Japan, were also added in this quarter.

Read next: Google to Withdraw Political Ads in EU Under Upcoming Transparency Standards
by Arooj Ahmed via Digital Information World

Google to Withdraw Political Ads in EU Under Upcoming Transparency Standards

Next year, Google plans to stop showing political ads to European Union users. The new EU rules—meant to promote transparency in political ads and protect voters—bring complex challenges for platforms like Google, making it hard to identify and manage ads across diverse regions. Google's concern is that the new regulations go too far, casting a broad net over what qualifies as a political ad, leading to operational uncertainty.

Google's strategy is to step back from political advertising in Europe ahead of these rules, which take effect in October 2025. Google’s decision isn't new; it has paused political ads in places like France, Canada, and Brazil when regulations lacked clear guidelines.

• Also read: A New Survey Finds Executives and Employees Think About AI Implementation Differently

The upcoming EU rules demand that political ads be labeled, showing who paid, how much they paid, and which elections or referendums they’re linked to. Advertisers also need user consent for targeting, and personal data of minors or sensitive information can’t be used. YouTube, too, must comply, banning paid political promotions under the EU's standards.

Back in February, Google had flagged the upcoming rule’s ambiguity, worried that unclear definitions might disrupt their ad recommendations. Google's current response is to adapt early and keep evaluating its approach.

Image: DIW-Aigen

Read next: 

• Millions Still Use Basic Passwords Despite Rising Data Breach Risks

• Study: Major Companies Lag in Reporting Phishing Scams Using Their Brand Names

by Asim BN via Digital Information World

A New Survey Finds Executives and Employees Think About AI Implementation Differently

According to a survey of 17,000 desk workers worldwide by Slack, employees and executives have different thoughts when it comes to using AI and its adoption in workplaces. 97% of the executives say that they are planning to invest on AI implementation, and 97% said that they want to do it as soon as possible. In the past month, the percentage of Americans who were happy to work with AI dropped from 45% to 36%. The UK, France and Japan also showed a decline in interest related to AI implementation.

This shows that even though many business leaders and executives are investing more and more on AI, employees do not seem too enthusiastic about it. On the other hand, there was some increase in AI usage in the US. According to the estimations, the usage of AI increased from 32% to 33% between March and August 2024. This is a slow growth rate, but it is still something.

76% of the employees surveyed also said that they are feeling an urgency to learn AI and it is making them anxious and less productive. But there are also employees who say that technology helps them with administrative tasks (87%), core work projects (80%) and creative work (81%). But they think that using AI is going to increase their workload.

Top three reasons why workers do not want to use AI are because they think that they are cheating (47%), they fear that other people will see them as lazy (46%) and fear that others will see them as less competent (46%).

Read next: Beyond Simple Math, AI Hits a Wall—FrontierMath Shows Where It’s Stuck
by Arooj Ahmed via Digital Information World

Meta's Threads Might Get Ads As Early As The Beginning of Next Year

Meta’s social media app Threads could be launching ads sooner than later.

The company just shared how it’s got major plans to introduce ads to the latest platform as early as 2025. This is certainly a shock considering how its parent firm Meta announced previously that this wouldn’t be the case.

The news comes to us thanks to a report published by The Information that spoke about a possible release in January 2025. This means Meta can monetize on the popular service than what may have been originally planned by Meta’s executives.

This year in August, the platform struck 200 million users which CEO Mark Zuckerberg referred to as a major milestone. He even hailed the platform to be the firm’s next billion-user service. This meant it would certainly be making loads of revenue through it but efforts on this end wouldn’t come too soon and would instead be a multi-year effort.

All of the latest products by Meta take time between shipping and scaling them into profitable businesses, Zuckerberg adds. The firm’s latest earnings call had CFO Susan Li adding how Threads might not be a major driver for 2025 as far as revenue was concerned.

As per reports from The Information, the slow rollout for ads on Threads will begin in January which is so soon. It’s also unclear if this effort will expand soon or not. The number one priority is to design consumer value first. For now, Meta’s spokesperson says it has no form of revenue generation or monetization on the app.

These newly reported plans for Meta showcase just how fast the service continues to grow with time. 275M people are working each month and seeing more than one million people signing up each day says so much, Zuckerberg highlights. Therefore, all of this makes the app one of the biggest alternatives for Musk’s X platform that really came into existence in the past few years.

Meanwhile, another famous app that offers features like Twitter is Bluesky whose user figures similarly saw a major rise recently. It went on to include one million new people in the past week, as confirmed by the tech giant on Tuesday. This is certainly not as much as what Threads could accomplish with 15M users but it’s definitely getting somewhere.

Similar to Threads, it has no advertising and the firm hopes to experiment with all features linked to subscriptions.

Image: DIW-Aigen

Read next:

• Google Gears Up For The Festive Season With Trending Picks For Gifts 2024 In Its ‘Holiday 100’

• Is YouTube Helping or Hurting Autism Awareness? New Study Reveals Surprising Results!

• Experience, Industry, and Company Size Drive Social Media Managers’ Salaries, New Survey Finds
by Dr. Hura Anwar via Digital Information World