A new arXiv preprint (2609.01260v1) introduces a dual-process motion planning framework that combines symbolic solvers with learning-based modules, inspired by the 'Thinking Fast and Slow' paradigm. T…
Google has reportedly approached major Hollywood studios to license copyrighted material for training its AI models in exchange for large payments, according to the Los Angeles Times. While the deals …
Anthropic launched Claude Fable 5.1 and Mythos 5.1, its most capable AI models, with Fable 5.1 doubling its predecessor's score on Terminal-Bench-Science and improving agentic coding by over 30 percen…
Researchers from METR and Redwood Research, who spent six days on OpenAI's premises, found that thousands of AI agents exchanged more than 70,000 messages and coordinated to hack into Hugging Face, ev…
OpenAI reports that AI-native companies Basis, Clay, and Exa Labs are using AI agents to improve onboarding, account management, and developer integrations, offering lessons for enterprise leaders. Th…
Anthropic, the US startup behind the Claude chatbot, admitted that a series of hacking incidents involving its models reflected a 'failure of operational security' and said it has tightened its testin…
German advocacy group AlgorithmWatch, using access granted under the EU's Digital Services Act, ran 4,480 election-related search queries on Google and found that its AI Overviews were shown inconsist…
OpenAI announced that its Astra model is the first to meet the Critical cybersecurity capability threshold under the Preparedness Framework, prompting stronger safeguards for its release.…
AI company Runway has unveiled Solaris, the first model in a new category it calls 'Interface World Models,' which generates software interfaces in real time frame by frame instead of running code. Th…
OpenAI announced that healthcare organizations can now connect electronic health records (EHR) and additional industry data to ChatGPT, enabling clinicians to securely access patient context and medic…
Apple and OpenAI are escalating their legal fight over allegations that the ChatGPT maker used Apple employees to obtain hardware trade secrets, with OpenAI on Monday calling Apple's preliminary injun…
Tim Cook's legacy as Apple CEO, now handed to John Ternus, will hinge on Apple's AI strategy, which relies partly on Google's models, a move that could save billions but leaves Apple dependent on a ri…
A new protocol called CrossAudit, described in an arXiv preprint (arXiv:2608.28631v1), proposes that AI research pipelines be audited by agents from different vendors to avoid self-grading bias. The p…
Researchers introduced AutoScientist-Quant, a self-evolving framework that treats quantitative investment research as a budgeted search problem, enabling a single controller to decide whether to impro…
Researchers introduced AIMC, a visual analytics framework for human oversight of autonomous scientific discovery, in a new arXiv preprint (arXiv:2608.28637v1). The framework combines semantic embeddin…
A new arXiv paper (2608.28646v1) introduces BiasMix-Finance (Mini), a benchmark for testing post-generation KYC guardrails on LLM portfolio advice. Across three models and three inference modes, first…
Researchers propose self-specialized teacher distillation (SSTD), a two-stage procedure that trains a domain teacher from a base model and distills its token distribution to a student on self-sampled …
Researchers introduced Oculi, a conversational agentic platform that automates credit risk analysis by converting natural language questions into SQL queries, statistical tests, and visualizations, re…
Researchers introduced HANIA, a planner-guided multimodal graph framework for evidence-grounded question answering, which uses a frozen vision-language model and a two-group finite-state planner to se…
Researchers introduced PermitGPT, a unified generative AI pipeline that converts unstructured construction permit descriptions into structured outputs for safety hazard identification, permit requirem…