AI News, Jul 12, 2026: Meta disables AI tool that generated images of Instagram users

Today's stories highlight a shift toward realism, as tech giants roll back privacy-invasive features and face the financial realities of the AI boom.

1. 01
   Study finds LinkedIn has the highest rate of AI-generated postsThe Decoder

   An analysis by Pangram found that 41% of long-form posts on LinkedIn are entirely AI-generated, making it the leading platform for automated content.

   Why it matters

   Helps users understand why their professional feeds feel increasingly generic, as over 40% of long-form posts are now flagged as AI-written.
2. 02
   Anthropic adds built-in web browser to Claude CodeThe Decoder

   Anthropic updated Claude Code with an integrated browser, allowing the AI assistant to interact with external web pages while requiring user approval for transactions.

   Why it matters

   Developers can now let the AI read, click, and type on external websites directly from their coding environment, speeding up web-based tasks.
3. 03
   S&P Global downgrades Oracle credit rating over OpenAI riskThe Decoder

   S&P Global cut Oracle's credit rating to one notch above junk status, citing the company's heavy financial exposure to OpenAI's massive data center contracts.

   Why it matters

   Highlights the financial risk of the AI infrastructure boom, showing how heavily cloud providers rely on a single startup's survival.
4. 04
   Meta disables AI tool that generated images of Instagram usersThe Decoder

   Meta has discontinued its Muse Image feature following backlash over allowing users to generate AI images of public Instagram accounts without consent.

   Why it matters

   Protects user privacy by stopping others from creating AI-generated images of you using your public Instagram photos without consent.
5. 05
   Anthropic report shows Claude Cowork is mostly used for mundane tasksThe Decoder

   An analysis of 1.2 million Claude Cowork sessions revealed that half of all usage is dedicated to routine business processes and text creation.

   Why it matters

   Shows that AI's primary value in the workplace is automating tedious administrative chores like status reports rather than replacing creative jobs.
6. 06
   OpenAI CEO pivots to predicting AI will create more jobs than it cutsThe Decoder

   OpenAI CEO Sam Altman stated he is now confident that AI is a net job creator, reversing his previous predictions of widespread professional disruption.

   Why it matters

   Signals a shift in how AI leaders talk about employment, moving away from warnings of mass layoffs to reassure workers and regulators.
7. 07
   Researchers use structured memory to help AI agents win complex gamesThe Decoder

   The AgenticSTS project successfully trained AI agents to win the game Slay the Spire 2 by replacing growing chat logs with a five-layer structured memory system.

   Why it matters

   Reduces the amount of data AI models need to process, lowering computing costs and making agents much more effective at long-term planning.