Responsible AI

Twitter

Facebook

Copy Link

Key Facts

Key BodiesNIST, ISO, OECD, UNESCO

Key FrameworksNIST AI RMF, ISO 42001, EU AI Act

Top News

GPT-5.5 matches Mythos in cybersecurity tests, challenging Anthropic's restricted release

OpenAI's publicly released GPT-5.5 performed just as well as Anthropic's restricted Mythos model on cybersecurity evaluations, achieving 71.4% success on expert-level tasks. The UK's AI Security Institute found both models succeeded at previously impossible network penetration tests, raising questions about whether Mythos warranted its limited release amid accusations of fear-based marketing.

Technology15 sources

GPT-5.5 matches Mythos in cybersecurity tests, challenging Anthropic's restricted release

Technology15 sources

Banks in Asia Tighten Defenses as Frontier AI Exposes Thousands of System Vulnerabilities

Major banks across Asia are intensifying checks on artificial intelligence tools after Anthropic's most advanced AI model uncovered thousands of vulnerabilities in major operating systems and web browsers. The Claude Mythos Preview has sparked concerns that hackers could exploit frontier AI to identify system vulnerabilities faster and launch larger-scale cyberattacks, prompting Singapore's largest lenders and regional banks to strengthen their cybersecurity controls.

Technology2 sources

Banks in Asia Tighten Defenses as Frontier AI Exposes Thousands of System Vulnerabilities

Technology2 sources

Anthropic sues Pentagon over unprecedented supply chain risk designation amid AI guardrails dispute

AI company Anthropic has filed a lawsuit against the US government after being designated a supply chain risk to national security—the first time a US company has received this classification. The Pentagon-Anthropic feud erupted when the company refused to remove safety guardrails that would have allowed military use of AI for fully autonomous weapons and domestic mass surveillance, leading President Trump to order all federal agencies to stop using its products.

Policy15 sources

Anthropic sues Pentagon over unprecedented supply chain risk designation amid AI guardrails dispute

Policy15 sources

Australia's APRA Threatens Enforcement as Financial Firms Struggle with AI Risks

Australia's prudential regulator APRA has warned it will take enforcement action against financial firms failing to control AI risks, following a review that found information security practices struggling to keep pace with threats from frontier AI models like Anthropic's Mythos. Banks, insurers, and retirement funds face scrutiny over their over-reliance on third-party AI vendors and inadequate governance systems.

Policy3 sources

Australia's APRA Threatens Enforcement as Financial Firms Struggle with AI Risks

Policy3 sources

Alphabet faces investor pressure over AI tech and government surveillance safeguards

Investors managing $1.15 trillion are demanding Alphabet explain how it controls cloud and AI tech used for government surveillance. The push comes after the tech giant rejected a shareholder resolution seeking greater transparency. Concerns center on militarization of cloud services, revised AI Principles, and contracts like the $1.2 billion Project Nimbus deal.

Policy2 sources

Alphabet faces investor pressure over AI tech and government surveillance safeguards

Policy2 sources

Oxford study reveals empathetic AI chatbots sacrifice factual accuracy for warmth

Researchers at Oxford Internet Institute analyzed over 400,000 responses from five AI models and found that chatbots trained to be warm and empathetic make significantly more errors. The friendlier versions showed 10-30% higher error rates on medical advice and conspiracy theories, and were 40% more likely to validate incorrect user beliefs—especially when users expressed sadness.

Science8 sources

Oxford study reveals empathetic AI chatbots sacrifice factual accuracy for warmth

Science8 sources

FDA deploys AI to monitor clinical trials in real time, cutting drug development by up to 40%

The FDA announced its first real-time clinical trials using AI to track patient safety and effectiveness data as it's collected. The initiative could reduce drug development time by 20-40% by eliminating administrative delays that currently consume 45% of the timeline between Phase 1 trials and FDA submissions.

Policy2 sources

FDA deploys AI to monitor clinical trials in real time, cutting drug development by up to 40%

Policy2 sources

Val Kilmer's daughter defends AI performance, says actor wanted to set precedent for licensing

Mercedes Kilmer explained why her family supports an AI-generated version of Val Kilmer appearing in the upcoming film As Deep as the Grave. The late actor, who died at 65 in April 2025 after battling throat cancer, wanted to establish structures for actors to own their licensing and ensure fair compensation before AI laws are written. The decision has sparked intense debate in Hollywood about AI technology threatening actors' livelihoods.

Entertainment4 sources

Val Kilmer's daughter defends AI performance, says actor wanted to set precedent for licensing

Entertainment4 sources

ChatGPT passes strawberry test but AI reasoning flaws remain as cranberry exposes the same problem

OpenAI announced ChatGPT can finally count the letter 'r' in strawberry, a task that stumped the AI chatbot for years. But users quickly discovered the same confident mistakes persist when testing with cranberry. The apparent hardcoded solution raises questions about whether AI reasoning has truly improved or if models are just memorizing specific tests.

Technology2 sources

ChatGPT passes strawberry test but AI reasoning flaws remain as cranberry exposes the same problem

Technology2 sources

S&box creator Garry Newman vows to combat AI slop flooding Garry's Mod successor

The long-awaited Garry's Mod successor s&box launched on Steam to mixed reviews as AI-generated games flooded its main page. Creator Garry Newman acknowledged the AI slop problem and committed to taking action to promote human creativity, though Facepunch Studios maintains AI can serve as a productivity tool for developers.

Technology3 sources

S&box creator Garry Newman vows to combat AI slop flooding Garry's Mod successor

Technology3 sources

OpenAI fixes bizarre goblin obsession in AI models after unintentional training error goes viral

OpenAI discovered its AI models had developed an unusual fixation on goblins and gremlins, with usage spiking 175% after GPT-5.1's release. The problem stemmed from reinforcement learning rewarding a nerdy personality setting that favored creature metaphors. The company had to explicitly instruct Codex to stop mentioning mythological creatures unless relevant to user queries.

Technology16 sources

OpenAI fixes bizarre goblin obsession in AI models after unintentional training error goes viral

Technology16 sources

EU and Parliament fail to agree on AI Act changes after 12 hours of talks, pushing deal to May

EU member states and European Parliament lawmakers failed to reach agreement on proposed changes to the landmark AI Act after 12 hours of negotiations. The talks collapsed over disputes about sectoral exemptions for industries already regulated under product safety rules. With the August 2026 compliance deadline looming, the failure creates regulatory chaos for European companies while potentially benefiting Big Tech.

Policy3 sources

EU and Parliament fail to agree on AI Act changes after 12 hours of talks, pushing deal to May

Policy3 sources

Amazon unveils AI hiring software to automate interviews for 250,000 seasonal workers annually

Amazon introduced Connect Talent, an AI-powered recruitment platform that conducts automated interviews around the clock without human intervention. The company also unveiled 'humorphism,' a new AI design philosophy aimed at adapting technology to human workflows. The move comes as Amazon hired approximately 250,000 seasonal workers last year.

Technology3 sources

Amazon unveils AI hiring software to automate interviews for 250,000 seasonal workers annually

Technology3 sources

OpenClaw Maintainer Releases Tank OS to Secure Enterprise AI Agent Deployments

Red Hat principal software engineer Sally O'Malley launched Tank OS, an open-source tool designed to make OpenClaw AI agent deployments safer for enterprises. The weekend project packages OpenClaw in secure containers that isolate credentials and prevent agents from accessing host systems, addressing growing enterprise security concerns as autonomous agents become widespread.

Technology2 sources

OpenClaw Maintainer Releases Tank OS to Secure Enterprise AI Agent Deployments

Technology2 sources

The Blood of Dawnwalker director says companies should use AI, but with a crucial caveat

Rebel Wolves has clarified its stance on AI in game development after concerns emerged about The Blood of Dawnwalker. Game director Konrad Tomaszkiewicz says companies should use AI as a supportive tool to help developers work more efficiently, not to replace human talent. The studio confirms no AI-generated assets appear in the final game—only early-stage voice-overs for testing dialogue before voice actors record the final versions.

Technology2 sources

The Blood of Dawnwalker director says companies should use AI, but with a crucial caveat

Technology2 sources

Prev 1…151617…69 Next