Microsoft patches Copilot flaw after eight-month delay — AI Weekly, 8–21 August 2026

8–21 August 2026 — the AI news that mattered this week, with a link to every source.

This week saw fixes for long-standing security gaps, new consumer-facing AI tools, and fresh research into protein design and traffic management. Governments and investors also moved to shape the next wave of AI infrastructure.

Five things to know

1. Microsoft patches Copilot flaw after eight-month delay The patched vulnerability could silently steal enterprise data, making it a priority for security teams.
   Microsoft finally patches critical one-click Copilot vulnerability, almost eight months after learning of it — CSO Online
2. OpenAI launches ChatGPT for Teens A tailored model for teenagers signals the first wave of age-specific consumer AI products.
   OpenAI launches ChatGPT for Teens, promising a more age-appropriate chatbot — BNN Bloomberg
3. Claude accelerates protein design research AI-driven protein engineering could shorten drug discovery timelines and reduce costs.
   How Claude is accelerating protein design and analytical chemistry — Anthropic
4. AWS launches agentic data pipeline automation Automating data engineering could cut weeks-long onboarding to hours, lowering operational overhead.
   Agentic Data Operations Platform (ADOP): Data engineering into hours — AWS Machine Learning
5. Agentic AI diagnoses aircraft IFEC issues Real-time diagnostics in aviation show how agentic systems can reduce downtime and maintenance costs.
   Accelerating aircraft IFEC diagnostics with agentic AI on AWS — AWS Machine Learning

Industry

Microsoft closed a critical Copilot flaw after months of exposure, while OpenAI’s expanded chain-of-thought monitoring raised costs for frontier models. AWS also advanced agentic automation for data pipelines and aircraft diagnostics.

• Microsoft finally patches critical one-click Copilot vulnerability, almost eight months after learning of it — CSO Online
• OpenAI — The Register
• Agentic Data Operations Platform (ADOP): Data engineering into hours — AWS Machine Learning
• Govern AI agent tool access with Amazon Bedrock AgentCore Gateway — AWS Machine Learning
• Reduce RAG costs on Amazon Bedrock with query-aware compression — AWS Machine Learning
• Accelerating aircraft IFEC diagnostics with agentic AI on AWS — AWS Machine Learning

Products

OpenAI launched ChatGPT for Teens, Google tuned its Discover feed with AI, and UPDF emerged as a lightweight PDF editor for the agentic era. Liquid AI also released draft models for faster decoding.

• OpenAI launches ChatGPT for Teens, promising a more age-appropriate chatbot — BNN Bloomberg
• OpenAI is gaining on Anthropic with business users, new data indicates — TechCrunch
• Google Discover is getting an AI chatbot-tuned feed — The Verge
• Introducing cross-Region inference for OpenAI GPT-5.6 models on Amazon Bedrock — AWS Machine Learning
• Meet UPDF: A Lightweight Adobe Alternative Built for the Agentic Era — MarkTechPost
• Liquid AI Releases LFM2.5-DSpark Draft Models That Deliver Up to 3.18x Faster Decoding Without Changing Model Outputs — MarkTechPost

Research

Claude’s protein design work and AI-identified drug candidates for childhood dementia showed how AI can accelerate scientific discovery. Traffic-light AI trials and greener ammonia production also featured.

• How Claude is accelerating protein design and analytical chemistry — Anthropic
• AI helps identify drugs that may slow childhood dementia — Times of India
• ET World Leaders Forum 2026: The heartbeat of India lives in Tier-2, Tier-3, rural areas, says Google India MD Preeti Lobana — The Economic Times
• Artificial intelligence powered traffic lights are being trialled, but will it ease traffic congestion? — ABC (Australian Broadcasting Corporation)
• Paving the way for greener ammonia production — MIT News
• A third of web pages published since ChatGPT’s launch show signs of AI authorship, study finds — TechCrunch

Policy

CISA flagged an actively exploited Ray flaw, Greater Manchester rejected a Palantir health contract, and India’s PM called for AI governance with human oversight. Anthropic also plans to change data retention policies.

• CISA Flags Actively Exploited Ray Flaw That Can Trigger Browser-Based RCE — The Hacker News
• The Single English County Saying No to Palantir — WIRED
• Anthropic plans to change data retention policy for advanced AI — The Star
• Runlayer, Rippling drop lawsuits — but the brouhaha is still a cautionary tale for founders — TechCrunch
• PM Modi Calls For Effective Use Of AI In Governance, Says ‘Human Judgement’ Must Be Ensured — Times Now

Funding

South Korea announced a chip windfall fund for AI and youth investment, while Relativity Networks raised $22 million for faster fibre technology. OpenAI’s slowed pace raised questions about its IPO plans.

• South Korea plans chip windfall fund to back youth, AI investment — Reuters
• It’s Greg Brockman’s OpenAI now — The Verge
• OpenAI hit the brakes. Now what? — The Verge
• VentureBeat names Rob Strechay as its first Lead Analyst, expanding its enterprise AI research push — VentureBeat
• Relativity Networks raises $22 million to bring a faster kind of fiber to data centers — TechCrunch

Worth watching

• Hugging Face Journal Club: Training AI Scientists to Replicate Research — Hugging Face
• What is an omni-model? — NVIDIA Developer

Every headline above links to the publisher that reported it. We summarise and link; we do not republish.