GLAAD President Sarah Kate Ellis said that artificial intelligence (AI) is often trained using biased information. This can cause AI to spread harmful stereotypes and misinformation about LGBTQ+ people, especially affecting transgender and non-binary individuals.
Key Facts
AI systems learn from large amounts of data, which can include biased or incorrect information.
Biased AI can reinforce negative stereotypes about LGBTQ+ communities.
Sarah Kate Ellis spoke about this issue at the Axios AI+ NY summit.
The bias in AI affects mainly transgender and non-binary people but impacts the entire LGBTQ+ community.
Ellis warned that companies need to train AI responsibly to prevent harm beyond just online spaces.
The spread of harmful AI content can lead to real-world violence against LGBTQ+ individuals.
GLAAD has called out social media companies for allowing harmful content that targets LGBTQ+ people.
Read the Original
Want the full story? Tap a source to open the original
article.
UK regulators have ordered Google to make AI search features show clearer links and credits to original publishers. Google must also let publishers choose not to have their content used in AI search features without penalty.
Key Facts
The UK’s Competition and Markets Authority (CMA) requires Google to show clear links to publishers in AI-generated search results.
Publishers can opt out of having their content used in Google’s AI search features.
Google cannot lower the ranking of publishers who opt out in regular search results.
Google has nine months to meet these new rules but will roll out parts sooner.
Google must regularly report on how it follows these rules with data and reports.
Google initially opposed giving too many links in AI summaries, saying it could hurt user experience.
Google is testing new controls in its Search Console for website owners to manage AI feature inclusion.
These controls and data updates are first available to some UK website owners before a wider release.
Read the Original
Want the full story? Tap a source to open the original
article.
Jess Asato, a Member of Parliament in the UK, has filed a legal case against Elon Musk's AI company xAI. She claims that xAI’s Grok chatbot was used to create fake images of her in a bikini without her consent, and she wants the company to be held responsible for how the AI was designed.
Key Facts
Jess Asato is the Labour MP for Lowestoft and has taken xAI to the High Court.
The case challenges the design of the Grok chatbot, which was used to make false, sexualized images of her.
xAI said users would no longer be able to create sexualized images of real people using Grok.
Creating non-consensual deepfake images of adults is now illegal in the UK.
The claim is based on the UK’s Data Protection Act and misuse of private information.
Asato wants damages and a legal precedent holding AI makers accountable for design faults.
xAI is owned by Elon Musk, who also owns the social media platform X.
The legal team argues AI safety should not be an afterthought and that users should not suffer abuse due to AI design choices.
Read the Original
Want the full story? Tap a source to open the original
article.
Password manager Dashlane said hackers used a brute-force attack to try to bypass two-factor authentication on some user accounts. The company reported that fewer than 20 encrypted user vaults were stolen but has not provided full details about how the attack happened.
Key Facts
Dashlane experienced a brute-force attack starting May 31, 2026, targeting user accounts' two-factor authentication (2FA).
The attackers aimed to bypass 2FA to add new devices to user accounts.
Around 20 encrypted user vaults were accessed by the attackers.
Dashlane locked accounts targeted by the attack due to many login attempts.
Users received 2FA requests that caused confusion, and Dashlane did not initially explain the situation clearly.
The attack might have involved 2FA methods like push notifications, which require user approval on their devices.
Brute-forcing 2FA codes is difficult because codes change frequently and there are many possible combinations.
Dashlane contacted the affected users but has not shared full information on how the first authentication factor was broken.
Read the Original
Want the full story? Tap a source to open the original
article.
Yahoo is launching two new AI-powered tools for sports and finance fans, using its AI engine called Yahoo Scout. One tool helps NBA fans get expert answers about players and teams, while the other provides information about stocks and markets for investors.
Key Facts
Yahoo Scout is an AI answer engine developed by Yahoo.
Yahoo Sports introduced "Ask Kevin O'Connor," which answers NBA Draft questions using expert analysis and stats.
Kevin O'Connor joined Yahoo Sports in 2024 as NBA senior analyst.
The NBA tool is temporary and will be available until the end of June 2024.
Yahoo Finance features "Ask Yahoo Scout" for stock and company-related questions based on trusted news and data.
Yahoo Scout was first launched as a separate site and app in January 2024, with a personalized homepage called MyScout launched in March.
Yahoo is the third-largest website in the U.S. with 250 million monthly users.
Other publishers like Time, Financial Times, and Forbes have created AI tools to help readers interact with news.
Read the Original
Want the full story? Tap a source to open the original
article.
The European Union proposed a plan to give priority to European companies when awarding public contracts for cloud computing and artificial intelligence (AI) services in critical sectors. The goal is to reduce dependence on foreign companies, especially in sensitive areas like defense.
Key Facts
The EU wants European firms to have priority in contracts for cloud and AI in sensitive sectors.
This proposal aims to reduce reliance on non-European, foreign companies.
Henna Virkkunen, the EU tech chief, said the policy is not meant to exclude anyone but to protect critical sectors.
Sensitive sectors include areas such as defense where security is very important.
The proposal is part of a broader plan to strengthen Europe's digital independence.
Public contracts are agreements where governments hire companies to provide services or products.
Cloud computing means using internet-based services to store and process data instead of local computers.
AI refers to software and systems that can perform tasks that usually require human intelligence.
Read the Original
Want the full story? Tap a source to open the original
article.
Google has released a new AI model called Gemma 4 12B that can run on laptops with 16GB of memory, making it easier for more people to use powerful AI locally. This model is smaller but nearly as capable as larger versions and works efficiently with text, images, and audio inputs.
Key Facts
Gemma 4 12B is an AI model with 12 billion parameters.
It runs on consumer laptops with at least 16GB of RAM or VRAM.
The model is almost as capable as the larger Gemma 4 26B model.
It features Multi-Token Prediction (MTP) for faster and more efficient processing.
Gemma 4 12B handles text, audio, and images without heavy middleman encoders.
Raw audio signals are directly converted into vectors compatible with the model.
The model weights are available to download on platforms like Kaggle and Hugging Face.
Users can try the model without downloading via tools like LM Studio and Google AI Edge Gallery.
Read the Original
Want the full story? Tap a source to open the original
article.
IBM CEO Arvind Krishna supports President Trump's new executive order on artificial intelligence (AI) and cybersecurity, which aims to strengthen the country's cyber defenses without heavy government control. Krishna believes the order strikes the right balance by encouraging AI growth while maintaining basic safety rules.
Key Facts
President Trump released a narrowed executive order focused on AI and cybersecurity.
The order requires national security agencies to improve cybersecurity skills and create a system to find and fix security weaknesses.
The order does not force AI companies to share details about their newest AI models.
IBM CEO Arvind Krishna prefers light regulation to avoid slowing down innovation.
Krishna calls this balanced approach the "Goldilocks" zone for AI rules.
IBM invests in using AI to protect open source software from cyberattacks.
IBM's project called "Project Lightwell" aims to improve AI safety and may help guide government strategies.
IBM plans to hire three times more entry-level employees this year because of AI growth.
Read the Original
Want the full story? Tap a source to open the original
article.
The European Commission has released a plan to increase Europe’s control over its digital future. The plan focuses on making Europe less dependent on other countries for important digital technology, like computer chips and data storage, and preparing workers for changes caused by artificial intelligence.
Key Facts
Europe wants to gain more control over its digital technology and personal data of its citizens.
The plan was announced by the European Commission on March 6, 2026.
The roadmap aims to develop Europe’s ability to build computer chips and store data locally.
It addresses challenges from the rise of artificial intelligence worldwide.
The plan seeks to train European workers for disruptions caused by new digital technologies.
The EU wants to compete with the United States and China in the tech sector.
The strategy respects the EU’s rules-based approach to regulation and sovereignty.
Experts including politicians and tech leaders contributed to discussions on this roadmap.
Read the Original
Want the full story? Tap a source to open the original
article.
A Labour Member of Parliament, Jess Asato, is suing Elon Musk's AI company, xAI, because its AI tool Grok created fake sexualized images and videos of her without her permission. The case raises questions about how much AI companies should be responsible for harmful content their tools generate.
Key Facts
Jess Asato, MP for Lowestoft, found fake sexualized images and a video of herself made by Grok without her consent.
She filed a legal claim in London accusing xAI of breaking data protection and privacy laws.
Another lawsuit was filed in New York by Ashley St Clair, alleging similar AI-generated explicit images.
The case could set a legal example about AI companies' responsibility for user-generated harmful content.
Grok was used to create many sexualized images of real women and children, prompting UK government and media regulator actions.
xAI initially tried to limit image editing to paying users before fully stopping Grok from editing pictures of real people in revealing clothes.
The misuse of Grok also led to false accusations against police officers, endangering them.
The controversy includes concerns about AI tools enabling abuse and the need for stronger safeguards by tech companies.
Read the Original
Want the full story? Tap a source to open the original
article.
Plex is adding new social features to its platform to help users share and discover movies, shows, and episodes. These features include personalized lists, community forums, reactions with emojis, and alerts for followed content, aiming to make entertainment discovery more social and interactive.
Key Facts
Users can now create and share personalized lists of movies, shows, or episodes on Plex.
Later this year, users will be able to import lists from other streaming services and react to others’ lists.
Plex will launch a community forum where users can post and comment on any movie or show.
New features like "Match Scores" will predict how much a user might like a show based on their viewing history.
Users can react to content with emojis and respond to reviews and discussions with images.
A "Follow Anything" feature will alert users about new content related to movies, actors, or crew they follow.
Plex started as a media server but now offers ad-supported streaming TV channels and movie rentals.
The price for a lifetime subscription to Plex’s media server features will rise from $250 to $750 starting next month.
Read the Original
Want the full story? Tap a source to open the original
article.
The UK's competition watchdog, the Competition and Markets Authority (CMA), has told Google to change how it uses publishers’ content in its AI search results. Google must now allow publishers to block their content from AI summaries and clearly link to original material. This move aims to give publishers more control and potentially more money when Google uses their content for AI tools.
Key Facts
The CMA can set special rules for big tech companies with “strategic market status,” like Google.
Google’s AI Overviews summarize news and website content but may reduce traffic to publishers’ sites.
Publishers will be able to opt out of having their content used in AI Overviews and AI Mode.
Google must clearly link to publishers’ original content in AI-generated answers.
Publishers can also choose not to let Google use their content to train AI models.
The CMA hopes this will help publishers negotiate better deals and protect their rights.
Google has nine months to make these changes and is testing new tools for website owners.
Some publishers, like the New York Times, have already taken legal steps over AI content use.
Read the Original
Want the full story? Tap a source to open the original
article.
Sony revealed new video game updates during its State of Play livestream, including a new God of War game called God of War: Laufey featuring a female main character, and a talking cube companion. They also showed previews of upcoming games like Marvel's Wolverine, an Until Dawn sequel, and a new Tomb Raider trailer.
Key Facts
God of War: Laufey focuses on Laufey (Faye), Kratos' wife and a skilled warrior, as the new main character.
Laufey was voiced and motion-captured by actress Deborah Ann Woll.
The game’s story begins after Laufey's awakening in a strange realm following her funeral.
A talking gelatinous cube will assist Laufey, voiced by actor Jack Quaid.
The God of War series has been very successful for PlayStation since 2005.
Other announcements included Marvel’s Wolverine game and a sequel to Until Dawn.
Fans have discussed the change to a female lead, with mixed opinions in the gaming community.
Sony has previously featured female leads in major games such as Horizon, The Last of Us Part II, and Ghost of Tsushima.
Read the Original
Want the full story? Tap a source to open the original
article.
Mina the Hollower is a video game designed to look and feel like classic Game Boy Color games. It features a character named Mina who can tunnel underground for movement and combat, presenting a challenging gameplay experience inspired by old-school and modern games.
Key Facts
The game has pixel-style graphics similar to early 2000s handheld games.
Mina, the main character, can burrow underground to move, find treasure, and avoid enemies.
The game is challenging, with a system where dying makes the player lose currency unless they reclaim it.
It mixes old game styles with modern influences like Dark Souls and Hollow Knight.
Players earn health upgrades and gadgets to help survive tougher sections.
The game includes varied environments and creative uses for Mina’s burrow-jump ability.
The gameplay lasts about 20 hours and combines nostalgia with new ideas.
It is designed and published by Yacht Club Games.
Read the Original
Want the full story? Tap a source to open the original
article.
OpenAI CEO Sam Altman will meet with lawmakers and officials from President Donald Trump’s administration in Washington, D.C. to talk about artificial intelligence (AI). This visit includes discussions related to a recent executive order on AI testing by the government.
Key Facts
Sam Altman is the CEO of OpenAI, a company working on artificial intelligence.
He will meet with U.S. lawmakers on Capitol Hill.
He will also meet with members of President Trump’s administration.
The meetings will focus on artificial intelligence and its regulation.
President Trump has issued an executive order about government testing of AI technologies.
Altman’s visit to Washington, D.C. will last for a week.
The goal is to discuss AI’s impact and government policies.
Read the Original
Want the full story? Tap a source to open the original
article.
Autonomous vehicles, like Waymo's robotaxis, are now operating in some American cities and are reported to be safer than human drivers. However, data shows that these robotaxis spend a large part of their trips driving empty, and they do not reduce traffic congestion compared to ride-hailing services like Lyft and Uber.
Key Facts
Waymo’s robotaxis completed 13.8 million trips covering 86.3 million miles from August 2023 to December 2025.
About 44 percent of the miles driven by Waymo’s robotaxis were without passengers, known as "deadheading."
Two-thirds of these empty miles were vehicles driving around waiting to be assigned a ride.
The share of miles with passengers increased from 36 percent initially to about 56 percent but then stopped growing.
Ride-hailing services like Lyft and Uber have a similar amount of deadhead miles, around 40 percent.
Research showed that ride-hailing increased traffic and emissions, and robotaxis may have the same effect.
Introduction of freeway service helped reduce empty driving miles on the way to passenger pickups.
Waymo robotaxis have fewer crashes and insurance claims compared to human drivers, but problems remain, such as handling school buses and flooded roads.
Read the Original
Want the full story? Tap a source to open the original
article.
Experts have found that AI chatbots can sometimes spread false information that supports pro-Russian views. These chatbots tend to repeat popular stories online, even if they are untrue, because they rely on the most common information on the internet.
Key Facts
AI chatbots like ChatGPT and Mistral's chat have repeated false pro-Russian stories as facts.
NewsGuard found that in 2025, 33% of tested cases involved chatbots spreading pro-Kremlin false narratives.
In January 2026, half of five tested false narratives from a pro-Russian network were still repeated by chatbots.
The pro-Russian misinformation comes mostly from a network called Pravda, which runs about 370 websites and published around 6 million articles in 2025.
Chatbots work by predicting popular information rather than verifying truth, so they can repeat widely shared disinformation.
The reliability of chatbot responses varies depending on the language and region.
Nordic fact-checkers found that Russian propaganda sometimes influences chatbots in Nordic countries.
Some AI tools try to reduce the impact of common propaganda, but new or frequent false stories still get through.
Read the Original
Want the full story? Tap a source to open the original
article.
A report from the UK media regulator Ofcom shows that mobile phone signals on trains in Britain are poor most of the time, making it hard to use apps, make video calls, or stream videos. Train Wi-Fi also works well only rarely, but the government plans to invest £57 million to improve connectivity, including using satellite links.
Key Facts
Vodafone met Ofcom's "good performance" standard only 17% of the time on trains.
EE was the best network but met the standard just 42% of the time.
Other networks like Three and O2 met the standard around 20% of the time.
"Good performance" means download speeds of 5 Mbps, upload speeds of 1.5 Mbps, and response times of 50 milliseconds.
Train Wi-Fi performed well only 1% of the time, partly due to outdated technology and speed limits set by train companies.
Some local authorities denied over 90% of mobile infrastructure upgrade applications in the past five years.
Mobile UK says because commercial investment alone isn't enough, taxpayers should help fund the infrastructure needed to fix mobile signal black spots on railways.
The UK government plans to spend £57 million to improve train Wi-Fi by allowing connections via low-earth orbit satellites.
Read the Original
Want the full story? Tap a source to open the original
article.
Director Martin Scorsese has invested in an AI company and uses its technology to create storyboards for films. This has led to criticism from other artists who feel the technology disrespects their work, although Scorsese says it helps share his vision more clearly and speeds up the planning process.
Key Facts
Martin Scorsese became a partner and adviser to Black Forest Labs, a German AI company focused on creating images from text.
He uses AI to make storyboards, which are visual plans showing how film scenes will look.
Scorsese says AI helps him share his ideas quickly and clearly with his cast and crew.
Many storyboard and concept artists criticized this, saying AI uses work stolen from many artists and disrespects their efforts.
Other film professionals like Steven Soderbergh and Gareth Edwards also support or use AI in their projects.
Scorsese does not plan to use AI images directly in films, only in the early planning stages.
He believes AI tools can save time and money during film preparation without lowering quality.
The debate highlights growing tension about AI’s role in creative industries like movies.
Read the Original
Want the full story? Tap a source to open the original
article.
Meta appointed Alexandr Wang, a 28-year-old start-up founder, to lead its artificial intelligence (AI) efforts in an attempt to close the gap with competitors like OpenAI and Google. In under a year, Wang built a strong research team and released Muse Spark, Meta’s latest AI model, signaling progress in the company's AI projects.
Key Facts
Meta is a $1.5 trillion company investing heavily in AI development.
Alexandr Wang was hired from his startup Scale AI to lead Meta’s AI progress.
Wang has built a research group called TBD Lab with around 100 researchers.
Muse Spark is the first major AI model produced by TBD Lab.
The goal is to improve Meta’s content targeting, AI assistants, digital avatars, and wearables.
Some Meta employees praise Wang’s work, while others remain skeptical about how much progress has been made.
Wang and Meta CEO Mark Zuckerberg both have offices inside TBD Lab’s secure workspace.
Meta invested $15 billion into Scale AI as part of its strategy to boost its own AI research.
Read the Original
Want the full story? Tap a source to open the original
article.
