AI News, Oct 8, 2026: OpenAI revenue reportedly $20 billion lower than previous estimates

Today's stories highlight the real-world friction of AI adoption, from financial reality checks and security flaws to physical safety risks.

1. 01
   Anthropic adds video generation and live data dashboards to ClaudeThe Decoder

   Anthropic released two new beta features for Claude, enabling the generation of live dashboards from databases and animated videos from text prompts.

   Why it matters

   Users can now turn raw data and text prompts into interactive dashboards and animated explainer videos directly within the chat interface.
2. 02
   OpenAI revenue reportedly $20 billion lower than previous estimatesTechCrunch

   A new report claims OpenAI's annualized revenue is significantly lower than the previously projected $70 billion.

   Why it matters

   This massive downward revision suggests the commercial market for generative AI may be growing slower than hyper-inflated private valuations imply.
3. 03
   Google launches autonomous AI agents for Gemini business usersTechCrunch

   Google announced new agentic capabilities for Gemini Enterprise, allowing the AI to plan and execute tasks across various business applications.

   Why it matters

   Workers can now delegate multi-step tasks to AI assistants that operate in the background, use multiple models, and have their own corporate email addresses.
4. 04
   Mathematicians call for OpenAI boycott over flood of AI-generated papersThe Decoder

   The Association for Human Mathematics urged a boycott after OpenAI released over 700 AI-generated math manuscripts, three of which were quickly retracted due to errors.

   Why it matters

   The backlash highlights growing friction between academic researchers and AI labs over the quality, accuracy, and sheer volume of AI-generated research.
5. 05
   USA Today publisher sues OpenAI for $250 million over copyrightThe Verge AI

   Gannett, the publisher of USA Today and hundreds of local newspapers, filed a lawsuit seeking damages from OpenAI for allegedly copying its articles to train models.

   Why it matters

   The lawsuit intensifies the legal battle over whether AI companies must pay publishers to use their articles for training data.
6. 06
   Anthropic updates policy to ban abusive behavior toward ClaudeThe Verge AI

   Anthropic updated its usage policy to prohibit sustained and needless verbal abuse of its Claude model, alongside stricter bans on high-risk use cases.

   Why it matters

   The policy sets a precedent for treating AI models with basic decency, while also tightening rules around election interference and weapons development.
7. 07
   Security flaw allowed complete takeover of AWS AI agentsThe Decoder

   Security researchers at Zenity Labs discovered a vulnerability in Amazon's Bedrock AgentCore that allowed a single prompt to hijack all agents in an AWS account.

   Why it matters

   Companies building cloud-based AI agents must secure internal interfaces, as a single prompt could have compromised entire cloud accounts.
8. 08
   Teen rescued from mountain after using Claude for route planningThe Decoder

   A 16-year-old hiker near Vancouver had to be airlifted by helicopter after Anthropic's Claude chatbot mistakenly guided him down a route requiring professional climbing gear.

   Why it matters

   The incident serves as a stark warning that AI models frequently hallucinate physical details and should not be trusted for wilderness navigation.