Telling AI It’s an ‘Expert’ Might Actually Make It Worse, Study Finds
New research shows that telling AI models they are “experts” can reduce accuracy in tasks like coding and math, while improving safety and alignment outcomes.
Chain-of-thought prompting is a technique intended to improve performance on problems that benefit from intermediate reasoning, such as multistep arithmetic, logic, planning, or decomposition. A prompt may include worked examples or ask the model to analyze a problem before answering. The visible explanation can help users inspect the approach, but it is not guaranteed to be a faithful record of the model’s internal computation and may still contain confident mistakes. For sensitive applications, the result should be checked against tools, evidence, or known answers. Modern reasoning systems may also perform hidden intermediate computation while returning only a concise conclusion or a short, user-facing explanation.
New research shows that telling AI models they are “experts” can reduce accuracy in tasks like coding and math, while improving safety and alignment outcomes.
Meta introduced four new in-house MTIA chips designed for AI training and inference as the company accelerates data center expansion. The chips aim to improve performance and reduce reliance on external hardware suppliers.
Google has rolled out Canvas in AI Mode to all U.S. users in English, enabling people to create documents, dashboards, and interactive tools directly within Search. The feature provides a dynamic workspace for planning projects and building simple applications.
Andrej Karpathy argues AI coding agents now fundamentally change programming workflows, enabling long-running autonomous tasks and redefining software engineering.
MSCI introduces AI connectors enabling access to its data via MSCI ONE, ChatGPT, and Claude, debuting IndexAI Insights to deliver conversational index analytics powered by large language models.
Google is rolling out Gemini 3.1 Pro across consumer, developer, and enterprise products, touting major gains in reasoning performance and benchmark results.
India’s AI Impact Summit in New Delhi delivered more than $250 billion in investment commitments, spanning data centers, AI infrastructure, and global expansion plans.
Fei-Fei Li’s World Labs secured a $200 million investment from Autodesk as part of a broader $1 billion round, deepening collaboration on AI systems that generate and reason about 3D environments.
Mistral AI CEO Arthur Mensch said more than half of today’s enterprise software could be replaced by AI, intensifying investor concerns about the future of SaaS business models.
Apple is developing multiple AI-powered wearable devices, including smart glasses, an AI pendant, and camera-enabled AirPods, according to a new report from Bloomberg.