AI News — Tuesday, March 17, 2026

Merriam-Webster and Encyclopedia Britannica are suing OpenAI, alleging copyright infringement of their extensive linguistic and factual databases.

A lawyer involved in AI psychosis cases is raising alarms about the potential for mass casualty risks stemming from the misuse or unpredictable behavior of advanced AI systems.

Innate is developing home robots that can be programmed with the same ease as AI agents, aiming to make advanced robotics accessible for everyday tasks.

Anthropic has released Cowork, a Claude Desktop agent that enables users to interact with their local files and perform tasks without requiring any coding knowledge.

Nvidia's new DLSS 5 technology leverages generative AI to enhance photorealism in video games, with potential applications extending beyond the gaming industry.

OpenAI has released Symphony, an open-source orchestrator designed to manage and coordinate autonomous development agents, streamlining complex software engineering tasks.

This research paper analyzes prompt injection attacks as a form of 'role confusion' in large language models, offering new insights into AI security vulnerabilities.

Sourcetable introduces an AI-powered spreadsheet and data platform designed to simplify data analysis and management through intelligent automation.

Dropbase AI offers an open-source, prompt-based builder that allows users to create Python web applications quickly using natural language commands.

Nous Research has introduced NousCoder-14B, an open-source coding model positioned as a free alternative to proprietary coding assistants like Claude Code.

LMEB introduces a new benchmark specifically designed to evaluate the long-horizon memory capabilities of large language models, addressing a critical area for advanced AI.

The 'Cheers' research proposes a method to decouple patch details from semantic representations, enabling large language models to achieve unified multimodal comprehension and generation.

daVinci-Env presents an open-source framework for synthesizing software engineering environments at scale, facilitating the development and testing of AI agents for coding tasks.

This article argues that the AI era is shifting software development away from stateless designs towards context-aware systems, fundamentally changing architectural paradigms.

Vellum provides a comprehensive development platform specifically designed to help engineers build, test, and deploy applications powered by large language models.