News reporting focused on AI and machine learning, covering the companies behind these technologies, their real-world applications, and the ethical concerns they raise. This includes areas like generative AI (large language models, text-to-image and video), speech tech, and predictive analytics.
A new study from three European universities found that access to AI advice dramatically reduced people's ability to recognize their own ignorance. Accuracy dropped from 27% to 9%, while confidence rose from 30% to 76%. The researchers used deliberately flawed AI to show the effect is not about sensible delegation but about cognitive suppression.
Researchers at Google DeepMind have developed GenCeption, a model that repurposes a pre-trained video generator for classic computer vision tasks like depth estimation and segmentation. The system builds on an open-source video model from Alibaba and delivers results in a single forward pass, guided by text prompts. It trained on just a small set of synthetic videos and needs far less data than competing approaches. In benchmarks, GenCeption matches established specialized models and transfers its abilities to real-world footage and untrained categories like animals. The authors say this supports the contested idea that video generators can serve as the basis for universal world models in computer vision.
Moonshot AI's Kimi K3 model is reportedly close to matching top Western AI models, raising fresh doubts about U.S. export controls. The model's performance, achieved with limited computing power, has impressed researchers and sparked debate about the role of compute in AI capability. An OpenAI strategist warns against open-weight models, while experts note the model's high cost per task.
A new audit of 2.5 million biomedical papers found fabricated references have increased more than twelvefold since 2023. Researchers at Columbia University and other institutions published the study in The Lancet, linking the rise to widespread use of language models like ChatGPT. The fake citations are hard to detect and pose special risks because they appear frequently in review articles that inform clinical guidelines.
Apple released macOS Tahoe 26.5, addressing numerous security vulnerabilities including a kernel bug (CVE-2026-28952) discovered by Calif.io in collaboration with Claude and Anthropic. The update fixes issues from denial-of-service to privilege escalation, sandbox escapes, and memory corruption. Multiple researchers and organizations are credited.
Security researchers at PromptArmor have discovered that Microsoft Copilot Cowork is vulnerable to file exfiltration attacks through indirect prompt injection. The flaw exploits automatic approval for sending messages to the active user, allowing attackers to steal pre-authenticated download links to files in SharePoint and OneDrive.
Researchers at arXiv have identified a phenomenon they call 'constraint decay,' where large language model agents experience a sharp drop in performance as structural requirements accumulate in backend code generation tasks. The study tested agents across 80 greenfield tasks and 20 feature-implementation tasks spanning eight web frameworks, finding that agents succeed in minimal frameworks like Flask but struggle significantly in convention-heavy environments such as FastAPI and Django. Data-layer defects, including incorrect query composition and ORM runtime violations, were identified as the leading cause of errors.
California Governor Gavin Newsom signed the first executive order by a U.S. governor aimed at shielding workers from AI-driven job displacement. The order directs state agencies to work with researchers, unions, and the AI industry on strategies including subsidies for companies that keep employees, expanded retraining for office workers, and a review of universal basic capital. Newsom cited predictions that half of all office jobs could vanish within five years and criticized the tax system for rewarding automation over hiring.
Elon Musk's SpaceXAI has seen more than 50 researchers and engineers leave since February, following its merger. Key teams in coding, world models, and Grok voice have lost leaders, with rivals like Meta and Thinking Machine Labs hiring many. Concerns grow over pre-training team cuts and work culture.
Richard Socher has launched Recursive Superintelligence, a San Francisco startup emerging from stealth with $650 million in funding. The company aims to build a recursively self-improving AI that fixes its own weaknesses without human help. Socher and his team of top researchers discussed their open-endedness approach in a recent interview.
Former Google, Meta and OpenAI researchers have started Recursive Superintelligence with over $650 million in funding and a $4 billion valuation. The six-month-old firm aims to create AI that improves itself. Backed by investors like GV, Greycroft, Nvidia and AMD, it joins efforts to automate AI development amid recent advances in code-writing AI systems.
Researchers at Google DeepMind are creating an AI-enhanced mouse pointer that understands context and user intent across applications. The system follows four principles to make interactions more natural, such as maintaining workflow flow and turning pixels into interactive entities. Demos show pointing at elements to edit images or get directions, with integration starting in Chrome and Googlebook.