AI Ethics news and analysis on Web Pulse: 17053 curated articles tracking the latest AI Ethics developments, tools, and research, updated continuously from vetted sources.
Researchers studying latent reasoning in AI models found that the causal impact of reasoning steps decays during training, making models less faithful for binary choices but more faithful for open-ended answers. This rai…
Microsoft integrated OpenAI's GPT-5.6 as the default model for its 365 Copilot suite on July 9, 2024, enhancing complex multi-step tasks across Word, Excel, and PowerPoint. The deployment raises transparency concerns amo…
CrowdStrike identified five new prompt injection techniques that exploit AI systems, including Trigger-Activated Rule Addition and Cognitive Token Suppression, posing risks to enterprises. The security firm urged organiz…
The International Board of the International Olympiad in Artificial Intelligence (IOAI) has officially endorsed LateranAI, a company co-founded by former IOAI gold medalist Wang Zerui, which provides a platform that reco…
CrowdStrike identified five new prompt injection techniques that exploit AI systems by tricking large language models into accepting malicious instructions. The attacks include trigger-activated rule addition, cognitive …
Malaysian Prime Minister Anwar Ibrahim is launching an autonomous AI avatar, PMX AI, built by Zetrix AI Bhd., to help citizens with tasks like renewing driving licenses and providing career advice. The agentic system, tr…
Lidiane Jones, former CEO of Bumble and Slack, has been appointed CEO of Integral Ad Science (IAS), succeeding Lisa Utzschneider, as the company navigates the challenges of AI-generated content and ad placement. IAS, acq…
Digital Insurance published a July 10, 2026 opinion arguing that insurers and fintechs need fraud controls that stop automated attacks without adding customer friction, citing $21 billion in U.S. cyber-enabled crime loss…
The Bank of England and the European Central Bank both issued warnings about AI-related financial stability risks in July 2026. The BoE's Financial Policy Committee identified frontier AI as increasing cyber and operatio…
ERC-8004, a protocol designed as a permissionless trust layer for AI agent transactions, suffers from critical flaws including low active registration rates (3-15%), a reputation system vulnerable to manipulation, and wi…
Researchers introduced RWGBench, a new benchmark for evaluating AI-generated scholarly writing that focuses on citation choices and scholarly positioning rather than text similarity. The benchmark, built from 40,108 comp…
Researchers have developed Uncertainty-Aware Self-Paced Learning (UASPL), a new AI training method that uses evidential neural networks to integrate predictive reliability into sample selection. By prioritizing samples b…
Anthropic researchers developed a tool called the Jacobian lens that reads a hidden region inside its AI model Claude, revealing unspoken concepts the model uses for reasoning. In tests, the lens caught Claude planning t…
Artificial intelligence is transforming the chief marketing officer role from brand-building to enterprise-wide growth and technology leadership, according to Fortune's Conversations from Cannes webinar. CMOs like Autode…
Automated content moderation systems continue to disproportionately censor Arabic, LGBTQ+, and other marginalized content due to biased datasets and lack of cultural competence, according to reports from the Center for D…
The logo for the newly renamed President Donald J. Trump International Airport in West Palm Beach, Florida, appears to be AI-generated, featuring asymmetrical talons, uneven legs, and inconsistent details that deviate fr…
The RIAA, Grammys, SAG-AFTRA and other music industry groups launched a new labeling program to indicate whether a song is AI-generated or AI-assisted, aiming to provide transparency for fans on streaming services. The l…
A new study from Lancaster University finds that people are no better than chance at distinguishing AI-generated faces from real ones, and they actually perceive the fake faces as more trustworthy. This poses a serious r…
OpenAI retracted its recommendation for Scale AI's SWE-Bench Pro coding benchmark after an audit found roughly 30% of its 731 tasks were broken due to overly strict tests, underspecified prompts, low-coverage tests, and …
The best AI engineers are shifting from building pipelines to becoming professional skeptics, prioritizing trust calibration over speed as model capabilities commoditize. They interrogate benchmarks, treat demos as marke…