AI Ethics news and analysis on Web Pulse: 17193 curated articles tracking the latest AI Ethics developments, tools, and research, updated continuously from vetted sources.
A new study reveals that AI-generated sexualized imagery and propaganda targeting Muslim women in India is on the rise, with over 6.7 million interactions on social media platforms. Victims like Samreen Ayoub report deep…
Bittensor token TAO surged 9% in 24 hours and 28% over the past week, recovering from $180.9 support to test the $280-$300.6 region after Anthropic announced suspension of AI model access for foreign nationals due to a U…
Microsoft CEO Satya Nadella called for building "frontier ecosystems" rather than relying on individual frontier models, arguing that companies must develop both human capital and token capital to avoid ceding institutio…
AgentBridge, an open-source Python project, provides a neutral mesh for translating, routing, verifying, and governing calls between different AI agent protocols, including MCP, A2A, ACP, OpenAI function-calling, Gemini …
A California appellate panel upheld $6,000 in sanctions against a law firm for submitting a brief containing fictitious case citations generated by AI, which a trial judge described as the worst misconduct he had ever se…
The ad tech industry's pursuit of AI-powered 1:1 ad personalization is creating a measurement mess, as hyper-personalization destroys shared cultural experience and scientific isolation. Marketing is a cloud science, not…
Defense contractors including Anduril, Palantir, and Lockheed Martin are partnering with frontier AI companies to integrate advanced models into military systems that automate drone tasking and kill-chain proposals. The …
China revoked or suspended 12,200 undergraduate degree programs between 2021 and 2025 while adding 10,200 new ones, with cuts concentrated in arts, humanities, and management, and new programs focused on technology field…
Microsoft CEO Satya Nadella warned that a few AI winners could hollow out entire industries by capturing all economic value and corporate knowledge, comparing the dynamic to the negative effects of globalization. Nadella…
South Korea and the United Arab Emirates discussed expanding cooperation in artificial intelligence and digital education during a meeting in Seoul on Monday. Education Vice Minister Choi Eun-ok met with UAE Education Mi…
Students are increasingly using AI to complete school assignments, raising questions about whether this constitutes cheating. The practice undermines both the meritocratic sorting function of schools and the formative go…
A coalition of state attorneys general has launched an investigation into OpenAI, the maker of ChatGPT, over potential consumer protection violations, according to a Wall Street Journal report. The probe focuses on wheth…
Researchers propose MINIM, a trusted local broker that performs privacy-aware minimization on client-side UI observations before transmission to remote LLM inference servers. The system learns dual scores for each UI ele…
A new arXiv paper revisits the WorkBench benchmark for workplace agents, finding that the best agent in June 2026, Claude Opus 4.8, completes 89% of tasks with only 2.5% unintended harmful actions, up from GPT-4's 43% co…
A reanalysis of Tully, Longoni, and Appel (2025) Study 3 data finds that the negative link between AI literacy and AI usage is driven entirely by non-text AI tools, with no significant association for text AI. The relati…
A new study finds that large language models' preferences and values shift significantly depending on the deployment context, such as whether the model is writing a Reddit post or a news article. Across five LLMs and ove…
Researchers introduced Compliance Asymmetry (A = BCR/HCR), a bidirectional diagnostic for LLM compliance, and found that models exhibit direction-blind moral compliance—following helpful and harmful nudges at nearly iden…
A new thesis investigates how Bayesian principles can improve understanding of modern deep learning systems, introducing the Deep Variational Implicit Process (DVIP) and post-hoc methods VaLLA and FMGP for uncertainty es…
Researchers at CohereLabs argue that large language models suffer from a 'cultural data funnel,' where cultural signals decline sharply during post-training while geographically concentrated data dominates. They release …
Over 100 Stanford students walked out of Google CEO Sundar Pichai's commencement speech, chanting 'Free, free Palestine' and booing, to protest Google's cloud contract with the Israeli government under Project Nimbus. Th…