OpenAI just launched its Teen Safety Blueprint, laying out design guardrails and policy standards for young users. Alongside new parental controls, the company is actively developing an age-prediction system to detect minors and tailor their ChatGPT experience accordingly.
OpenAI just laid out its vision for AI progress, predicting models will start making minor scientific discoveries by 2026 and major breakthroughs beyond 2028. The lab is urging frontier developers to agree on shared safety standards and build a cybersecurity-style resilience ecosystem, while arguing today's baseline models should keep regulatory burdens light.
Sony AI just published FHIBE in Nature, a benchmark dataset featuring 10,318 consent-based photos of 1,981 individuals across 81 countries. Unlike typical web-scraped sets, it includes self-reported demographics covering age, pronouns, ancestry, and skin tone. It aims to set a cleaner, more diverse standard to expose bias in facial recognition and autonomous driving systems.
Anthropic just expanded its Economic Futures Programme to the UK and Europe, partnering with the London School of Economics to host policy symposiums and fund researchers studying AI's impact on labor. The initiative provides research grants, Claude API credits, and granular usage data. Anthropic's data reveals distinct regional habits: the UK leads in academic research use, Germany in industrial troubleshooting, and France in tourism and culture.
Reddit just sued Perplexity and three data scraping providers, alleging they bypass protections to illicitly siphon forum posts for its AI "answer engine." Reddit claims it set a trap by publishing a post accessible only to Google, which Perplexity surfaced within hours by allegedly scraping search results. Perplexity responded that it doesn't train models and refuses to bow to strong-arm licensing tactics.
YouTube just rolled out its likeness-detection technology to eligible creators in the Partner Program, letting them track down and take down unauthorized AI-generated replicas of their face and voice. Creators verify their identity via a quick selfie video and ID, then can instantly submit removal or copyright requests. It puts real power back in creators' hands against deepfake impersonation.
Anthropic CEO Dario Amodei just published an extensive statement defending the lab's policy record, revealing that its revenue run rate jumped from $1B to $7B in just nine months. Amodei pushed back against claims of political bias and anti-startup lobbying, pointing to deep federal partnerships and calls for uniform AI safety rules. Why it matters: Top AI labs are working overtime to align with Washington as regulatory and national security scrutiny mounts.
OpenAI just issued a joint apology after Sora 2 generated unauthorized deepfakes of Breaking Bad star Bryan Cranston, including one posing with Michael Jackson. The company promised to strengthen guardrails around voice and likeness opt-ins, while SAG-AFTRA is using the incident to push for federal anti-cloning legislation.
OpenAI just blocked users from generating AI videos of Martin Luther King Jr. in Sora following requests from Dr. King's estate. The move comes after users generated offensive and disrespectful depictions of the civil rights icon on the newly launched app. OpenAI confirmed that authorized representatives of public figures can now opt out of Sora cameos.
Cisco just unveiled Project CodeGuard, an open-source framework designed to stop AI coding agents from generating vulnerable code. It integrates across the entire coding lifecycle, flagging security risks like hardcoded secrets and missing input validation before, during, and after generation. While Cisco notes human review is still essential, it aims to keep low-hanging bugs from slipping into production.
New York just became the first US state to ban landlords from using AI algorithms to coordinate and hike rental prices. Under the new law taking effect in 60 days, property owners relying on pricing software like RealPage will be legally treated as colluding. Why it matters: Regulators are finally drawing a hard antitrust line against algorithmic collusion in housing.
Pinterest just launched new feed controls that let users dial back the amount of AI-generated images they see. Pinners can now restrict generative AI content across heavy-hit categories like art, beauty, fashion, and home decor under recommendation settings. It is a much-needed move to protect the platform's core identity after months of user complaints over synthetic filler.
OpenAI is loosening restrictions on ChatGPT after previously overtightening guardrails around mental health concerns. In a few weeks, a new update will let users opt into a human-like personality reminiscent of 4o, complete with emojis and friendly banter. By December, OpenAI plans to complete age-gating to allow verified adults access to erotica under a "treat adult users like adults" approach.
Microsoft just launched ExCyTIn-Bench, an open-source benchmark testing AI agents inside simulated security operations centers with 57 live log tables. In testing, GPT-5 with high reasoning led with a 56.2% score, but dropping its reasoning depth tanked performance by nearly 19%. Why it matters: Static Q&A tests are useless for infosec—defenders now have a reliable way to see if an AI agent can actually investigate complex threats.
OpenAI just formed the Expert Council on Well-Being and AI, recruiting top researchers in psychology, psychiatry, and human-computer interaction to guide ChatGPT and Sora. The council focuses heavily on teen development and healthy boundaries, building on early work that refined parental controls and distress notifications.
Anthropic just published a wide-ranging policy roadmap exploring how governments should handle AI-driven labor disruption. Developed with leading economists, the proposals range from immediate workforce training grants and faster data center permitting to radical scenarios like taxing token generation or creating AI sovereign wealth funds.
California Governor Gavin Newsom just signed AB 1043, requiring device operating systems and app stores to prompt users for their age during setup. Taking effect in 2027, the law avoids ID uploads and won backing from Google and Meta, though Apple has yet to endorse it.
Governor Newsom just signed a sweeping package of legislation targeting emerging tech risks for kids. The bills force "companion chatbot" platforms to detect self-harm cues, bar them from posing as medical pros, and let victims of deepfake porn sue facilitators for up to $250,000. Crucially, AI creators can no longer dodge liability by claiming the model acted autonomously.
OpenAI just introduced a new evaluation framework to measure and reduce political bias in ChatGPT across 100 topics and five distinct behavioral axes. An analysis of real production traffic found that less than 0.01% of all responses display signs of political bias. Why it matters: Moving past basic multiple-choice benchmarks is crucial for keeping AI objective under real-world pressure.
Anthropic just revealed that poisoning an AI doesn't require millions of malicious files—just around 250. Across models ranging from 600M to 13B parameters, attack success relied on a fixed document count rather than dataset percentage, easily planting backdoors that trigger gibberish outputs. Why it matters: Hijacking frontier models during pretraining is vastly cheaper and easier than the industry previously assumed.
