AI Transparency

Twitter

Facebook

Copy Link

Key Facts

Key MethodsModel Cards, Datasheets

Required ByEU AI Act

Top News

Guide Labs open-sources Steerling-8B, a new interpretable LLM that traces every token to source

San Francisco startup Guide Labs has open-sourced Steerling-8B, an 8-billion parameter language model built with a novel architecture that makes every generated token traceable to its training data origins. Founded by Julius Adebayo and Aya Abdelsalam Ismail, the company aims to solve AI's black box problem by engineering interpretability directly into the model rather than analyzing it post-hoc.

Technology2 sources

Guide Labs open-sources Steerling-8B, a new interpretable LLM that traces every token to source

Technology2 sources

Pinterest AI moderation flags human art while AI-generated content floods user feeds

Pinterest users report widespread issues with the platform's AI moderation system incorrectly flagging hand-drawn artwork as AI-modified while actual AI-generated content overwhelms their feeds. Artists say the faulty account moderation is banning users randomly and forcing them to spend hours appealing decisions, undermining their brands and wasting time they could spend creating art.

Technology2 sources

Pinterest AI moderation flags human art while AI-generated content floods user feeds

Technology2 sources

Sam Altman warns companies are using AI washing to mask layoffs as job cuts surge

OpenAI CEO Sam Altman called out companies for AI washing, claiming firms are falsely attributing layoffs to AI when broader economic conditions are the real culprit. While nearly 55,000 job cuts cited AI in 2025, research shows 90% of executives report no AI impact on employment, suggesting corporate excuses for downsizing rather than genuine automation-driven displacement.

Business3 sources

Sam Altman warns companies are using AI washing to mask layoffs as job cuts surge

Business3 sources

UN panel aims for human control of AI as Guterres calls for facts over fear

UN chief Antonio Guterres announced that the UN General Assembly has confirmed 40 members for the Independent International Scientific Panel on Artificial Intelligence, emphasizing the need to make human control of AI a technical reality. The panel's first report is expected in July, ahead of the UN Global Dialogue on AI Governance.

Policy2 sources

UN panel aims for human control of AI as Guterres calls for facts over fear

Policy2 sources

AI agents surge in capability but lack safety disclosures, MIT AI Agent Index reveals

MIT researchers analyzed 30 AI agents and found a troubling pattern: while developers eagerly showcase what their systems can do, only 19% disclose formal safety policies and fewer than 10% report external safety evaluations. The study exposes how agentic AI systems that can autonomously perform tasks like booking flights and managing workflows are advancing faster than their safety frameworks.

Science6 sources

AI agents surge in capability but lack safety disclosures, MIT AI Agent Index reveals

Science6 sources

Tom Cruise Brad Pitt AI fight scene sparks Hollywood debate over green-screen footage controversy

A viral AI video featuring Tom Cruise and Brad Pitt in a rooftop fight scene generated over 1.8 million views and triggered panic across Hollywood. But visual effects artist Aron Peterson uncovered evidence suggesting ByteDance's Seedance 2.0 used green-screen footage with stuntmen rather than creating the scene from a simple prompt alone, raising questions about AI video generation capabilities.

Entertainment2 sources

Tom Cruise Brad Pitt AI fight scene sparks Hollywood debate over green-screen footage controversy

Entertainment2 sources

Fidesz Sparks Outrage With AI-Generated Election Video Showing Soldier's Execution

Viktor Orban's Fidesz party released a controversial AI-generated campaign video depicting a Hungarian soldier's execution, drawing fierce condemnation from opposition leader Peter Magyar. The 33-second video, confirmed to use Google's AI models, highlights growing concerns about artificial intelligence in political campaigns ahead of Hungary's April 12 parliamentary election.

Entertainment2 sources

Fidesz Sparks Outrage With AI-Generated Election Video Showing Soldier's Execution

Entertainment2 sources

Google Gemini rolls out Lyria 3 to generate 30-second AI songs from text and images

Google has integrated DeepMind's Lyria 3 model into the Gemini app, enabling users worldwide to create 30-second music tracks from simple text or image prompts. The feature auto-generates lyrics and produces custom cover art, marking a significant expansion of AI music generation capabilities. While Google emphasizes copyright protections and SynthID watermarking, the move raises questions about artistic authenticity and the music industry's future.

Technology20 sources

Google Gemini rolls out Lyria 3 to generate 30-second AI songs from text and images

Technology20 sources

Bug exposed confidential emails to Copilot AI, bypassing Microsoft's data protection policies

Microsoft confirmed a coding bug allowed Copilot AI to access and summarize confidential emails for nearly a month, despite data loss prevention policies designed to shield sensitive information. The flaw, tracked as CW1226324, affected Microsoft 365 Copilot Chat users since January, raising concerns about AI tool security as organizations struggle to keep pace with rapid feature rollouts.

Technology16 sources

Bug exposed confidential emails to Copilot AI, bypassing Microsoft's data protection policies

Technology16 sources

Galgotias University removed from AI Summit after claiming Chinese robot dog as own creation

Galgotias University was expelled from India's flagship AI Summit after a professor claimed a commercially available Chinese-made robotic dog was developed by the university. The incident, which drew swift online backlash, cast an uncomfortable shadow over an event designed to position India as a global AI powerhouse and showcase Made in India innovation.

Entertainment17 sources

Galgotias University removed from AI Summit after claiming Chinese robot dog as own creation

Entertainment17 sources

Google reveals Glimmer design system for Android XR AI glasses ahead of 2026 launch

Google has detailed its Glimmer design language for Android XR AI glasses, introducing a physics-aware UI philosophy built specifically for transparent displays. The system prioritizes voice-first interaction, eye-tracking, and glanceable transient UI elements that appear only when needed, marking a departure from traditional Material Design principles.

Technology3 sources

Google reveals Glimmer design system for Android XR AI glasses ahead of 2026 launch

Technology3 sources

Experts warn no AI can detect deepfakes as India grapples with national security threat

At the India AI Impact Summit, cybersecurity experts revealed that no AI model can reliably detect sophisticated deepfakes, warning this could become a national security risk by 2029 elections. Meanwhile, MeitY scientists propose curbing virality of synthetic content and exploring content provenance standards like C2PA to build accountability frameworks.

Policy3 sources

Experts warn no AI can detect deepfakes as India grapples with national security threat

Policy3 sources

Samsung's AI-generated ads for Galaxy S26 camera features spark widespread backlash

Samsung is flooding its social media channels with AI-generated advertisements to promote the upcoming Galaxy S26 series, but the strategy is backfiring. Videos showcasing camera features contain obvious artificial elements like unnaturally weighted shopping bags and shifting cobblestones. Critics are calling it shameless AI slop, with some viewers threatening to abandon Samsung for competitors.

Entertainment2 sources

Samsung's AI-generated ads for Galaxy S26 camera features spark widespread backlash

Entertainment2 sources

Big Tech's AI Climate Claims Called Greenwashing as Evidence Falls Short

Big Tech companies like Google and Microsoft claim AI will help solve climate change, promising to cut global emissions by 5-10% by 2030. But a new report analyzing 154 statements finds that 36% of AI climate claims lack any evidence, while companies conflate traditional AI with energy-intensive generative AI—a tactic researchers call greenwashing.

Policy5 sources

Big Tech's AI Climate Claims Called Greenwashing as Evidence Falls Short

Policy5 sources

15 Global Tech Giants Launch Trusted Tech Alliance to Set New Standards for AI and Infrastructure

At the Munich Security Conference, 15 companies from 10 countries—including AWS, Microsoft, Google Cloud, Anthropic, Ericsson, and Nokia—announced the Trusted Tech Alliance. The initiative establishes shared principles for trusted technology across connectivity, cloud infrastructure, semiconductors, software, and AI. Members commit to transparency, security, and data protection standards regardless of nationality.

Policy2 sources

15 Global Tech Giants Launch Trusted Tech Alliance to Set New Standards for AI and Infrastructure

Policy2 sources

Prev 1…222324…38 Next