AI News, Sep 23, 2026: Anthropic's AI discovers gene-editing enzyme system

Today's updates highlight a shift toward cheaper, more integrated AI agents alongside growing regulatory and security scrutiny.

1. 01
   California bills target data center energy and water transparencyThe Verge AI

   California Governor Gavin Newsom signed legislation requiring data centers to disclose more information about their electricity and water consumption.

   Why it matters

   Local communities will gain better visibility into how the massive data centers powering AI impact their utility bills and resources.
2. 02
   Anthropic's AI discovers gene-editing enzyme systemThe Verge AI

   Anthropic announced that its Claude AI model autonomously discovered a new enzyme system similar to the gene-editing tool Crispr.

   Why it matters

   This demonstrates that AI can autonomously make significant scientific discoveries, potentially accelerating the development of new medical treatments.
3. 03
   ChatGPT Voice gets access to email, calendar, and SlackThe Decoder

   OpenAI updated ChatGPT Voice with access to external plugins, allowing users to interact with services like Slack and email using natural speech.

   Why it matters

   You can now manage daily tasks, send emails, and coordinate schedules entirely through spoken conversation with your AI assistant.
4. 04
   YouTube adds AI script coaching and editing to Creator StudioThe Decoder

   YouTube announced a suite of new AI tools for its Creator Studio, including a storytelling assistant, Gemini-powered editing, and live translation.

   Why it matters

   Video creators can use built-in AI to refine scripts, generate smart thumbnails, and translate live streams into other languages.
5. 05
   US lawmakers propose prison time for building superintelligent AIThe Verge AI

   Senator Bernie Sanders and Representative Greg Casar introduced a bill to ban the development of artificial superintelligence, threatening violators with prison.

   Why it matters

   If passed, this strict legislation would criminalize the development of highly advanced AI systems deemed a threat to humanity.
6. 06
   Anthropic releases Opus 5.5 modelAI Business

   Anthropic launched its new flagship model, Claude Opus 5.5, offering improved performance but maintaining its premium pricing structure.

   Why it matters

   Users get access to a more capable flagship model, though it remains priced at a premium compared to cheaper rivals.
7. 07
   Meta patches critical security flaw in Muse AI assistantWIRED AI

   Meta issued a fix for a zero-day vulnerability in its Muse AI assistant that could have allowed attackers to execute arbitrary code on victims' Macs.

   Why it matters

   The vulnerability highlights how malicious actors can exploit AI assistants to gain full control over a user's computer.
8. 08
   OpenAI releases cheaper GPT-6 Sol and Luna modelsMarkTechPost

   OpenAI released GPT-6 Sol and GPT-6 Luna, two lower-cost models with improved prompt caching that reduce API costs by up to 50 percent.

   Why it matters

   Lower API prices make it much cheaper for developers to build and run complex AI agents that require long-running conversations.