📖 2 min read
Friday brought a packed news cycle across the AI world — from new safety features in ChatGPT to fresh funding for autonomous agents and a deepening alliance between Anthropic and SpaceX. Here’s what mattered in the last 12 hours.
OpenAI rolls out “Trusted Contact” in ChatGPT
OpenAI launched a new Trusted Contact feature that lets ChatGPT users designate a friend or family member to be alerted if a conversation suggests possible self-harm. The opt-in tool arrives as AI companies face mounting scrutiny over chatbot behavior in sensitive moments.
📧 Want more like this? Get our free The Ultimate AI Tool Database: 200+ Tools Rated & Ranked — Downloaded 5,000+ times
The same product update wave also introduced advanced voice models in the API and an early test of ads inside ChatGPT — a notable shift for a product that has so far stayed ad-free. Source: OpenAI
Anthropic and SpaceX strike a compute deal
Anthropic announced a compute partnership with SpaceX that hints at space-based AI infrastructure down the road, while expanding usage limits for Claude Code and the Claude API. Reports also point to Google preparing a follow-on investment in Anthropic worth up to $40B — pairing cash with cloud capacity to keep up with demand for Claude’s enterprise products.
Anthropic separately rolled out financial agents aimed at investment research and analysis workflows, doubling down on the agentic-finance trend. For a side-by-side of how Claude stacks up against the rest of the field, see our roundup on AiToolCrush.com.
📧 Want more like this? Get our free The Ultimate AI Tool Database: 200+ Tools Rated & Ranked — Downloaded 5,000+ times
AWS Bedrock AgentCore goes GA
AWS introduced a managed harness in Amazon Bedrock AgentCore that deploys autonomous AI agents in just three API calls, alongside a new CLI, persistent filesystem, and reusable skills. It’s a clear bet that the value in AI is migrating from raw models — increasingly commoditized — to the orchestration layer that lets agents plan, execute, and self-correct across systems.
ServiceNow followed suit at its Knowledge 2026 conference, announcing a broad rollout of autonomous agents across IT, security, customer workflows, supply chain, and device management.
Fazeshift raises $22M for autonomous finance agents
San Francisco–based Fazeshift closed a $17M Series A (bringing total funding to $22M) for AI agents that run end-to-end accounts receivable workflows — chasing invoices, reconciling payments, and closing books with minimal human touch.
📧 Want more like this? Get our free The Ultimate AI Tool Database: 200+ Tools Rated & Ranked — Downloaded 5,000+ times
It’s another data point for the thesis that vertical agent startups are eating into traditional SaaS budgets. If you’re tracking the AI-finance crossover, our latest model picks on BetOnAI.net include several tools competing in the same lane.
Google makes Gemini File Search multimodal
Google upgraded the Gemini API File Search to handle multimodal inputs, making retrieval-augmented generation and agent-building noticeably more practical for teams building over mixed image, document, and video corpora. Combined with this week’s GPT-5.5 Instant push from OpenAI, the gap between frontier labs keeps narrowing on production-grade RAG.
That’s the wrap for today. Check back tomorrow morning for the next edition of Today in AI.