Introducing Gemini 3.8 Live and 3.8 Live Extended Thinking
Google DeepMind has announced Gemini 3.8 Live along with a 3.8 Live Extended Thinking variant.
Week of September 14, 2026
September 14, 2026 – September 20, 2026 · 18 storiesOpenAI dominated the week with a string of customer case studies built around GPT-6 Astra, from Perplexity to Hex. The company also launched Astra for Law, a dedicated offering for the legal sector, and introduced new advertising formats including Sponsored Agents. On the safety side, it published a framework for reporting model misalignment alongside six reports, plus a youth safety blueprint for Australia. Google DeepMind announced Gemini 3.8 Live and an Extended Thinking variant, while Google released new findings from its AI & Economy Atlas.
Model releases
Agents & harnesses
Perplexity trusts GPT-6 Astra with end-to-end systems
OpenAI describes how Perplexity relies on GPT-6 Astra to draft communications, modify software and watch over production systems. According to the case study, the team needs to supervise the model far less often than with previous models.
Your Agent Aced the Task. Will It Do It Again?
A Hugging Face blog post examines whether AI agents that succeed at a task do so reliably on repeated attempts.
Hex turns complex analysis into visual reports with GPT-6 Astra
Data platform Hex uses GPT-6 Astra in its data agents to turn answers into interactive visualizations. OpenAI presents it as a customer case study.
Research
New insights from Google’s AI & Economy ATLAS
Google has published new findings from its AI & Economy Atlas, a project tracking the economic effects of artificial intelligence. The post presents version 1.0 of the atlas.
AI for Societal Impact
Google has assembled a collection showing how experts and local leaders apply recent AI advances. The stated goal is to spread the opportunities created by AI more widely.
Building AI to accelerate science and improve lives
Google presents its work on AI aimed at speeding up science and improving people’s lives, illustrated with examples from education and healthcare.
AI for everyone in every language
Google highlights its efforts to make AI available in many languages so that more people can use it.
How workers are unlocking new ways of working
New economic research from OpenAI looks at how workers use AI for tasks outside their usual roles. It also identifies which of these new activities become a regular part of their jobs.
New experts join Google’s AI & Economy team
Google has added new experts to its AI & Economy research program, which studies the economic impact of artificial intelligence.
Products & apps
How Fyxer built an AI executive assistant people trust
Fyxer combines OpenAI models with fine-tuning, memory and user feedback to sort inboxes and write emails that match each user’s style. The case study presents how the startup built trust in its AI assistant.
How to connect AI usage to business value
OpenAI explains how analytics in ChatGPT Work and Codex help companies track AI usage and spending. The tools are meant to reveal training needs and link adoption to business results.
Reimagining advertising with AI
OpenAI is introducing new AI-based advertising formats, including Sponsored Agents and tools for marketers. The company also announced integrations with HubSpot and Shopify.
Helping older adults use AI in everyday life
OpenAI and AARP will run free, hands-on ChatGPT workshops for 1,000 older adults in 10 U.S. cities. The program aims to teach practical and safe use of AI.
Introducing Astra for Law
OpenAI has launched an offering for the legal sector with frontier models, custom workflows for law firms and connections to legal data sources. It includes controls designed for confidential client work.
How Cooley is accelerating IPO work with ChatGPT
Law firm Cooley built a tool called GO Public with ChatGPT Work to support IPO projects. It helps lawyers spot issues earlier in the process.
Safety & security
Our framework for reporting model misalignment
OpenAI has published a framework for tracking, investigating and disclosing cases where its models behave in misaligned ways. It released six reports on unexpected or concerning model behavior at the same time.
Introducing the Australian Youth Safety Blueprint
OpenAI has released a six-pillar roadmap for Australia meant to make AI experiences safer for young people. The blueprint also aims to give them more control over how they use AI.