Lab

AI news flash: AT&T is working with OpenAI to make legal knowledge easier to access and use. According to the source, AT&T’s internal LegalEdge team is rethinking how legal matters are taken in (intake) and trying to scale legal judgment, so more people can get consistent legal help when they need it instead of relying entirely on a small number of legal staff handling cases one by one. However, the English summary provided is only two sentences long…

AI news flash: OpenAI explains how it works with distribution partners to keep safety standards consistent as access broadens. According to the source summary, the company works with these partners to keep safeguards aligned across channels while supporting responsible use, so more users and organizations can get access. Another focus is clarifying what the Usage Policies expect from each party, namely…

AI news flash: OpenAI is teaming up with PerturbAI and Scripps Research to use distributed AI agents to analyze the huge volume of data generated by millions of brain cells and surface promising research leads. Cell-level biological data is massive and complex, and manually reviewing it item by item can’t cover everything, so the team splits the analysis across many AI agents running at the same time…

AI news flash: OpenAI described how it built the ChatGPT ad system, with a large language model (LLM) driving the entire ad stack. The system tackles three problems. First, understanding what users actually intend in a conversation instead of just matching keywords, so ads fit the context of the chat. Second, keeping ads clearly separated from ChatGPT’s answers so ad content can’t influence or leak into the model’s responses, which protects the independence and credibility of…

AI news flash: OpenAI and Google Cloud announced a partnership that brings OpenAI’s frontier models and Codex workflows into Gemini’s Model Garden, available through Google Cloud Marketplace. This means Google Cloud’s enterprise customers can use OpenAI models without leaving their existing cloud environment …

AI news flash: Terminal X’s Eval Agent is an evaluation mechanism for checking the quality of AI output. It scores each output item by item against a rubric of 54 criteria, with a focus on whether the output strays from the fund’s own investment mandate, meaning whether the AI-generated content violates the fund’s set investment scope and limits…

AI News Flash: OpenAI has been working on strengthening ChatGPT’s performance on low-resource programming languages, rolling out a new round of evaluations focused on software engineering tasks, SQLite-related operations, and COBOL code migration capabilities. Compared to mainstream languages like Python and JavaScript, these languages have long been a weak spot for large language models due to scarce training data. This update adds dedicated benchmarks to examine the model’s real-world accuracy and reliability when writing, debugging, and porting code in these languages, helping identify capability gaps for targeted improvement.

AI News Brief: OpenAI has launched the ‘Rosalind Academic Plan,’ offering sponsored access to GPT-Rosalind and related tools to eligible academic labs to help advance life sciences research. The original source summary is fairly brief, stating only that the plan’s core positioning is to open sponsored slots to academic institutions so research teams can use GPT-Rosalind for work in the life sciences, without providing specifics on eligibility thresholds, the scope of sponsorship, which companion tools are included, or the plan’s launch timeline.

AI News Brief: OpenAI launched an ‘AI for Educators’ Teacher’s Day feature for the Brazilian Portuguese market, highlighting how OpenAI’s Study Mode helps teachers support students through active, step-by-step learning. The original summary is brief and doesn’t include specifics on Study Mode’s mechanics, technical details, or real-world data — see the source link for more.

AI News Brief: Block is partnering with OpenAI to build an AI-powered security system, with the core goal of dramatically speeding up threat detection and remediation. Block’s internal security team is integrating AI into daily threat investigation workflows to help quickly identify potential attack behavior and analyze anomaly patterns, and once a vulnerability or threat is found, AI helps propose a remediation fix — the whole process can be completed in minutes.