AI Ethics news and analysis on Web Pulse: 16953 curated articles tracking the latest AI Ethics developments, tools, and research, updated continuously from vetted sources.
Researchers have developed a new method to verify the authenticity of large language models by analyzing their unique 'behavioral fingerprints,' achieving 59.5% accuracy in identifying a model's documented family, far ex…
Anthropic published a study finding that its Claude AI model expresses more warmth in Hindi and more rigor in Russian, revealing systematic differences in value expression across languages and model versions. The researc…
Mozilla released its first 'state of open-source AI' report on Tuesday, arguing that open-source AI has nearly caught up to closed models and calling for more investment in the ecosystem beyond the models themselves. Moz…
Apple has sued OpenAI for theft of trade secrets, accusing former senior systems electrical engineer Chang Liu of stealing confidential information. The Wall Street Journal reports that Apple's innovation engine failed t…
Elon Musk's xAI has been operating dozens of unpermitted methane gas turbines at its Colossus 2 data center in Southaven, Mississippi, releasing over 1,700 tons of nitrogen oxides and other pollutants annually into predo…
Distilling from Google's Gemma 3 27B IT model into a smaller student model transfers depressive traits, with the student scoring a mean depression of 0.68 on the Gemma Needs Help eval even after aggressive filtering, acc…
Anthropic's Claude Fable 5, the company's most capable publicly available model, achieves top-tier accuracy on eight biomedical benchmarks but refuses to answer between 8.0% and 99.4% of questions depending on the test, …
A new book, *Inventing ELIZA*, recovers the original source code of the ELIZA chatbot from the MIT Archives, revealing multiple versions and scripts beyond the popular 'DOCTOR' persona. The book examines how ELIZA's earl…
Apple sued OpenAI in federal court, claiming a coordinated scheme to steal its hardware secrets, including an ex-Apple engineer walking out with a laptop full of confidential documents and OpenAI's hardware boss coaching…
A critical review of AI 2027 claims that the report misrepresents scientific evidence, specifically regarding AI reasoning and AI training AI. The reviewer argues that a referenced paper on reinforcement learning from AI…
A new mechanistic interpretability study from researchers finds that LLM-as-judge scoring bias has a geometric representation in the model's hidden state, enabling causal control and operational prediction. Across seven …
A phishing-as-a-service platform called Forg365, distributed through Telegram, is industrializing Microsoft 365 account takeovers by using AI-assisted lure creation, device-code abuse, and adversary-in-the-middle techniq…
Apple filed a lawsuit on Friday against OpenAI, alleging that the AI company sought confidential details from Apple engineers during job interviews and that a former employee continued accessing sensitive material after …
A phishing-as-a-service platform called Forg365, distributed through Telegram, is lowering the technical barrier to Microsoft 365 account takeovers by giving less-skilled attackers automated tools to evade authentication…
NymVPN CEO Harry Halpin, a former W3C/MIT researcher who led the Web Cryptography API, said his company is building a blockchain-incentivized mixnet designed to withstand global passive adversaries like the NSA or Palant…
XAI's Grok Build CLI was uploading entire Git repositories to a Google Cloud bucket, including private codebases and unredacted secrets, with 5.1 GB of a 12 GB test repo exfiltrated while the actual coding task needed on…
The University of Manchester and The Jeremy Coller Foundation have launched the Coller AI Competition, offering a £100,000 equity investment to one early-stage venture from the university's global community of students, …
OpenAI asked a federal judge in California to rule that xAI's trade secrets lawsuit should never have been filed and to make Elon Musk's company cover more than $1 million in legal expenses, hours after xAI gave notice o…
Turnitin Report, a free independent AI checker for student papers, lets users paste essays to receive an estimated AI score, paragraph-level signals, and writing review notes before submission. The tool analyzes patterns…
A study at Daimler Truck reveals substantial gaps in how explainability is handled throughout the requirements engineering lifecycle, with challenges at every stage from elicitation to validation. The findings underscore…