AI Update, August 8, 2025: AI News and Views From the Past Week

Catch up on select AI news and developments from the past two weeks (in no particular order):

OpenAI launches GPT-5, aiming for AGI and enhanced usability. OpenAI has released GPT-5, the successor to GPT-4, emphasizing usability and safety improvements. CEO Sam Altman described it as offering PhD-level expertise on demand and a step toward artificial general intelligence. GPT-5 is now available with usage limits for free ChatGPT users and is being integrated into Microsoft's Copilot. While technical improvements over GPT-4 are modest, experts see it as a foundational reset for future innovations.

Anthropic revokes OpenAI's Claude API access over terms violations. Anthropic has cut off OpenAI's API access to Claude, citing violations of its commercial terms of service. OpenAI allegedly used Claude's capabilities for internal benchmarking via developer tools, bypassing standard interfaces. While Anthropic maintains it may still allow access for safety evaluations, the move highlights growing tensions and competitive tactics in the AI industry. OpenAI expressed disappointment, noting its own API remains open to Anthropic.

Apple builds new team to develop AI answer engine. Apple is reportedly forming a team to develop a proprietary AI-powered "answer engine" capable of pulling information from across the Web, potentially embedded into Siri, Safari, or as a standalone app. The project, named Answers, Knowledge, and Information, marks a significant move toward AI-driven search, possibly in response to Google's antitrust issues. Job listings suggest Apple is hiring experts in search engine development. While Apple has partnered with OpenAI to integrate ChatGPT, it is now pursuing a more direct AI approach.