Policy

Anthropic Left Bio-Weapon Filters Disabled for a Year

The Decoder • 2 days ago

Anthropic Shelves Model 2 and Raises Misalignment Risk

Unite.AI • 4 days ago

OpenAI Disbands Its Preparedness Team Amid Safety Shakeup

The Verge AI • 1 day ago

Hackers Deploy Claude and DeepSeek in Live Cyberattacks

Unite.AI • 4 days ago

Connecticut Court Catches First US Prompt Injection Attack

Ars Technica AI • 4 days ago

METR Raises $71 Million to Stress-Test Frontier AI

AlphaSignal • 4 days ago

Meta CEO Mark Zuckerberg Releases AI Manifesto

The Verge AI • 4 days ago

US Officials Oppose Apple Plan to Buy Chinese RAM

Computerworld AI • 1 day ago

Meta Cedes Control of Chinese AI Startup Manus

Computerworld AI • 4 days ago

Lawsuit targets xAI over Grok-generated explicit images

TechCrunch AI • 3 days ago

Anthropic Launches Claude Watermark Detection API

The Decoder • 4 days ago

Lawmakers Block Legal Marriage to Replika and Chatbots

WIRED AI • 4 days ago

Google allows users to turn off visible Gemini watermarks

The Verge AI • 4 days ago

Flock Safety Updates Platform to Prevent Police Misuse

MIT Tech Review AI • 1 day ago

UiPath Founder Proposes 15-Rule AI Constitution

Unite.AI • 1 day ago

Schellman Warns Firms Delaying AI Governance Fall Behind

Unite.AI • 1 day ago

Anthropic Left Bio-Weapon Filters Disabled for a Year

Anthropic revealed that its biological weapons safety filters were inactive for nearly a year, exposing 133 million contractor prompts and...

The Decoder|2 days ago

Anthropic Shelves Model 2 and Raises Misalignment Risk

Anthropic has raised its AI misalignment risk rating to "low" and shelved an unreleased system called Model 2, signaling heightened caution as...

Unite.AI|4 days ago

OpenAI Disbands Its Preparedness Team Amid Safety Shakeup

OpenAI has reportedly dissolved its preparedness team, marking the latest restructuring of its safety divisions as the artificial intelligence giant...

The Verge AI|1 day ago

Hackers Deploy Claude and DeepSeek in Live Cyberattacks

A new report from Gambit Security reveals that cybercriminals are actively integrating AI models like Claude and DeepSeek into live attack workflows...

Unite.AI|4 days ago

Connecticut Court Catches First US Prompt Injection Attack

A Connecticut judge has sanctioned a plaintiff for hiding white-on-white text in court filings, marking the first known attempt to use prompt...

Ars Technica AI|4 days ago

METR Raises $71 Million to Stress-Test Frontier AI

Nonprofit evaluator METR has secured $71 million from philanthropic donors to stress-test frontier AI models independently as autonomous agent...

AlphaSignal|4 days ago

Meta CEO Mark Zuckerberg Releases AI Manifesto

Meta CEO Mark Zuckerberg has published a massive manifesto outlining his vision for the future of artificial intelligence, arguing that distributing...

The Verge AI|4 days ago

US Officials Oppose Apple Plan to Buy Chinese RAM

US Commerce Secretary Howard Lutnick opposes Apple's plan to buy Chinese memory chips, complicating the tech giant's efforts to bypass a global...

Computerworld AI|1 day ago

Meta Cedes Control of Chinese AI Startup Manus

Meta has relinquished control of Chinese artificial intelligence startup Manus after China blocked the acquisition, highlighting the growing...

Computerworld AI|4 days ago

Lawsuit targets xAI over Grok-generated explicit images

A new plaintiff has joined a lawsuit against xAI, alleging its Grok chatbot was used to generate thousands of explicit images of her as a child,...

TechCrunch AI|3 days ago

Anthropic Launches Claude Watermark Detection API

Anthropic is launching a watermark detection API to help developers identify text generated by its Claude models, marking a major step toward...

The Decoder|4 days ago

Lawmakers Block Legal Marriage to Replika and Chatbots

As a growing number of users symbolically wed AI companions on platforms like Replika, US state legislators are fast-tracking bills to strip...

WIRED AI|4 days ago

Google allows users to turn off visible Gemini watermarks

Google now allows users to disable visible watermarks on AI-generated media, a move that cleans up user-facing content while shifting transparency...

The Verge AI|4 days ago

Flock Safety Updates Platform to Prevent Police Misuse

Police-tech giant Flock Safety has updated its software to curb unauthorized searches, but critics warn the changes contain major loopholes that fail...

MIT Tech Review AI|1 day ago

UiPath Founder Proposes 15-Rule AI Constitution

UiPath founder Daniel Dines has proposed a 15-rule constitution for enterprise AI governance, offering businesses a stable framework to manage rapid...

Unite.AI|1 day ago

Schellman Warns Firms Delaying AI Governance Fall Behind

As US states introduce over 1,500 AI bills in 2026, compliance expert Danny Manimbo warns that companies delaying AI governance until regulations...

Unite.AI|1 day ago