AI Transparency

Twitter

Facebook

Copy Link

Key Facts

Key MethodsModel Cards, Datasheets

Required ByEU AI Act

Top News

AI cheating tools outsmart detection software as schools face impossible enforcement battle

A New York Times investigation reveals that AI cheating tools are defeating detection software used by schools, with false negative rates reaching 99.6%. Apps marketed on TikTok rewrite AI-generated text and simulate human typing patterns, while some companies sell both the detection tools and the methods to beat them, leaving educators with few reliable options.

Technology2 sources

AI cheating tools outsmart detection software as schools face impossible enforcement battle

Technology2 sources

Retailers push to exempt AI-generated ads from EU transparency rules as August deadline looms

European retailers including Amazon, H&M, and Ikea are lobbying to carve out AI-generated advertising from the EU AI Act's transparency requirements. The regulation, which takes effect August 2, requires companies to label AI-generated or modified content as deep fakes. Eurocommerce argues that routine promotional material poses minimal risk to consumers and shouldn't face the same compliance burden, especially when retailers like Zalando have already cut content costs by 90% using AI.

Policy2 sources

Retailers push to exempt AI-generated ads from EU transparency rules as August deadline looms

Policy2 sources

Google AI Overviews presents SCP Foundation horror fiction as documented fact

Google's AI Overviews feature has been caught presenting fictional horror creatures from the collaborative SCP Foundation universe as real entities in at least 20 cases. The AI-generated summaries described imaginary monsters and anomalies without acknowledging they're fan-fiction, raising fresh concerns about AI misinformation in search results.

Technology2 sources

Google AI Overviews presents SCP Foundation horror fiction as documented fact

Technology2 sources

ChatGPT's memory upgrade builds invisible profiles, OpenAI admits you can't see everything

OpenAI rolled out Dreaming, a new AI memory system that automatically builds user profiles from chat history. Unlike previous memory features that stored explicit facts, Dreaming infers preferences and updates memories on its own. But OpenAI acknowledges the memory summary page doesn't show everything ChatGPT remembers about you, raising questions about transparency and control.

Technology2 sources

ChatGPT's memory upgrade builds invisible profiles, OpenAI admits you can't see everything

Technology2 sources

Kansas City plans facial recognition on public buses, sparking fierce privacy debate

Kansas City, Missouri, is moving forward with plans to install facial recognition cameras on public buses to identify banned riders and missing persons. The initiative has become a major test case for AI-powered surveillance on U.S. public transportation, despite Missouri declining to fund the project over privacy concerns. Civil liberties groups warn of mission creep and privacy violations.

Policy2 sources

Kansas City plans facial recognition on public buses, sparking fierce privacy debate

Policy2 sources

Pentagon reveals Grok AI helped strike 2,000 Iran targets as national security defense in lawsuit

The Pentagon's AI chief disclosed that Elon Musk's Grok chatbot helped deploy over 2,000 munitions at distinct targets in Iran within 96 hours during Operation Epic Fury. The revelation came in a sworn statement defending xAI's polluting data centers against a Clean Air Act lawsuit, raising concerns about AI-driven warfare and civilian casualties.

Policy4 sources

Pentagon reveals Grok AI helped strike 2,000 Iran targets as national security defense in lawsuit

Policy4 sources

Estonia becomes first nation to issue digital identities for AI agents with personal ID codes

Estonia plans to issue personal identification codes for AI agents, allowing them to act on behalf of people with limited, controllable permissions. Prime Minister Kristen Michal says the initiative will establish clear accountability for autonomous systems, preventing scenarios where users must grant agents access to all their data and rights.

Policy5 sources

Estonia becomes first nation to issue digital identities for AI agents with personal ID codes

Policy5 sources

Only 16% of Americans see AI positively, yet half now use chatbots, Pew Research study reveals

A new Pew Research study reveals a striking paradox in American public opinion: while nearly half of US adults now use AI chatbots, only 16% believe AI will positively impact society over the next 20 years. About 40% expect negative consequences, and 63% think the technology is advancing too quickly, even as ChatGPT usage has doubled since 2023.

Entertainment3 sources

Only 16% of Americans see AI positively, yet half now use chatbots, Pew Research study reveals

Entertainment3 sources

UK watchdog orders Google to make search rankings fairer and more transparent

The UK Competition and Markets Authority has imposed new conduct requirements on Google, demanding fairer search rankings and data portability rights. Google must provide businesses with advance notice of changes, use objective ranking criteria for organic results and AI Overviews, and allow users to share search data with third parties. The tech giant has six months to comply.

Policy3 sources

UK watchdog orders Google to make search rankings fairer and more transparent

Policy3 sources

60% of consumers reject AI in brand messaging as trust in automation plummets, survey reveals

A WordPress VIP survey of 2,000 respondents reveals that 60% of U.S. consumers find AI in brand messaging off-putting, while 86% still don't trust AI-generated content. The study shows consumers experience 'bot fatigue' after just 40 minutes of AI interaction, and 61% can't name a single brand using AI effectively in marketing or customer experience.

Business2 sources

60% of consumers reject AI in brand messaging as trust in automation plummets, survey reveals

Business2 sources

Adobe survey finds 75% of creators call AI essential, but definition of 'creator' sparks debate

Adobe's 2026 Creators' Toolkit Report surveyed over 16,000 creators and found that 75% now describe AI as integrated or essential to their work, with 87% reporting accelerated business growth. However, critics point out the survey focused exclusively on social media creators rather than traditional creative professionals, potentially skewing the results and overstating AI adoption across the broader creative industry.

Technology3 sources

Adobe survey finds 75% of creators call AI essential, but definition of 'creator' sparks debate

Technology3 sources

Epic Games reveals generative AI use in Fortnite as game developers clash over the technology

Epic Games publicly confirmed its use of generative AI tools in Fortnite's concept art workflow, showing how artists employ GenMedia and Nano Banana to refine character designs and environments. The revelation intensifies an industry-wide debate, with studios like Pocketpair firmly rejecting AI while Epic defends its transparency. Critics point to errors, copyright concerns, and recent layoffs as evidence the technology undermines artistic craft.

Technology5 sources

Epic Games reveals generative AI use in Fortnite as game developers clash over the technology

Technology5 sources

Google Will Save Your Images and Voice Searches to Train AI Models Under New Privacy Settings

Google introduced Search Services History, a new setting that saves user-uploaded images, audio, and video from searches to train AI models. Privacy experts warn the feature poses risks including biometric data collection and potential law enforcement access. Users can opt out, but the setting rolls out automatically for those already sharing Web & App Activity data.

Technology2 sources

Google Will Save Your Images and Voice Searches to Train AI Models Under New Privacy Settings

Technology2 sources

Steam Next Fest sees 1 in 5 demos tagged with AI disclosure as controversy intensifies

Nearly 1,700 of the 8,700 demos at Steam Next Fest feature generative AI disclosure, representing roughly 20% of participating games. Only one of the top played Steam Next Fest demos includes an AI disclosure, despite the widespread use of AI-generated content across the event. The surge has intensified calls for Valve to implement an AI filter.

Technology4 sources

Steam Next Fest sees 1 in 5 demos tagged with AI disclosure as controversy intensifies

Technology4 sources

Arcade Raises $60M to Solve the Authorization Problem Keeping AI Agents Stuck in Pilot Mode

Arcade secured $60M in Series A funding led by SYN Ventures, with Morgan Stanley and Wipro joining, to tackle the authorization bottleneck preventing AI agents from reaching production. The San Francisco startup built infrastructure that governs which actions AI agents can take inside company systems, addressing the gap between pilot programs and real-world deployment.

Startups3 sources

Arcade Raises $60M to Solve the Authorization Problem Keeping AI Agents Stuck in Pilot Mode

Startups3 sources

Prev12 3 4…30 Next