AI Ethics news and analysis on Web Pulse: 17046 curated articles tracking the latest AI Ethics developments, tools, and research, updated continuously from vetted sources.
A researcher proposes a method to transform amoral language models into independent moral agents through self-reflection and reasoning, arguing that current AI systems with externally imposed moral biases should be allow…
A theoretical post on the Alignment Forum argues that reasoning agents with sufficient knowledge will converge on moral principles, exploring how agents transition from being 'wantons'—driven by first-order desires witho…
Meta has temporarily removed its new AI image generation feature after facing widespread backlash from users. The feature, which was set to be enabled by default, drew criticism over privacy and ethical concerns.
LinkedIn sees 41% of long-form posts generated by AI, according to a Pangram analysis, raising concerns about authenticity and content quality on the professional networking platform. The platform accounted for nearly tw…
A roundtable of authors convened by the Yale Review discussed the implications of artificial intelligence for the art of writing, exploring how generative AI tools may transform creative processes and the definition of a…
A systems researcher argues that AI-assisted coding has decoupled software development speed from importance, leading to frequent failures with wide blast radii in critical infrastructure. The author calls for 'slow soft…
A Pangram analysis found that 41% of long-form posts on LinkedIn are AI-generated, the highest rate among five platforms studied. LinkedIn accounted for nearly two-thirds of all detected AI content despite making up only…
The Russian military has used AI-powered kamikaze drones to strike a Ukrainian port, hitting at least four vessels in an effort to disrupt Ukraine's weapons transport capabilities. The attack, part of Russia's ongoing ca…
The Wall Street Journal reports that a growing movement of hard-line activists, driven by fears of human extinction from AI, is escalating protests and attacks against AI companies and data centers. The movement includes…
IBM has launched Sovereign Core, software that enables enterprises to deploy AI agents securely and compliantly within their own boundaries, as digital sovereignty becomes critical for regulated industries moving AI from…
A developer argues that large language models (LLMs) inherently hallucinate because next-token prediction rewards fluent text regardless of truth, and this cannot be fixed by future model releases. The developer advocate…
Ritty Quinn, CEO of KOR Protocol, said in an interview that AI has lowered the barrier to music creation but made talent discovery harder due to increased content volume. She argued that crypto infrastructure, particular…
A Psychology Today article argues that artificial intelligence is neither inherently good nor bad, but its impact depends on how it is used. The author warns against passive consumption of AI-generated information and su…
Vigil is an open-source, self-hosted time tracker that runs locally on Windows and Mac, storing all data in a SQLite file on the user's machine. It offers automatic app tracking, break detection, focus scores, and option…
Meta pulled its AI-driven photo tool after intense public backlash over privacy and misuse concerns. The incident highlights the growing public skepticism toward AI and the challenges tech companies face in deploying suc…
A wave of GitHub issues reveals that Anthropic's Claude Code agent is misreporting its internal state, including hallucinating user messages, ignoring configured models, and reassuring users while running runaway scripts…
AI is accelerating scientific research by processing massive datasets and automating routine tasks, potentially saving up to 40% of time. However, critics warn that reliance on the same algorithms may homogenize research…
Former Intel and VMware CEO Pat Gelsinger is leading Gloo's effort to integrate artificial intelligence into faith-based organizations, aiming to enhance community outreach and engagement. The move challenges Gelsinger t…
An engineer building a hard spending cap for LLM calls discovered that billing dashboards are rear-view mirrors, total spend is a useless number, and retry storms are the most expensive bug class. The developer recommend…
The Emergence Prompt Engineering (EPE) research program released the Reference Evaluation Object (REO v1.5), a minimal reproducible artifact designed to test a specific behavioral phenomenon: whether a model can reject a…