AI Feb. 2025 Wrap-up

Here is a summary of important AI developments in February 2025, with also some relevance to insurance:

AI Industry Developments

• OpenAI may launch AI-powered gadgets like AR glasses, smartwatches, and robots (TLDR AI, 06 Feb 2025).
• Mistral introduced a revamped ‘le Chat’ AI assistant with new iOS and Android apps and enhanced speed (The Rundown AI, 07 Feb 2025).
• Google Search is evolving towards a more AI-driven experience (StrictlyVC, 05 Feb 2025).
• Anthropic is set to release a new Claude AI model featuring hybrid reasoning (The Rundown AI, 14 Feb 2025).
• "Deep Research" functionalities offered by ChatGPT, Perplexity, and Anthropic:
  • ChatGPT: Autonomous research for complex tasks, generating detailed reports with citations. Analyzes text, images, and PDFs.
  • Perplexity AI: Generates structured reports based on autonomous real-time searches. Offers citation-backed AI search.
  • Anthropic: Capabilities to defend against jailbreaks and enhance AI safety, though implementation in a deep search application is unknown. Combines LLM capabilities with reasoning algorithms in a single system, but implementation in a deep search application is unknown .
• Jensen Huang's Vision for AI Advancement (Note: multiple sources, Jan & Feb)
  • Agentic AI and Robotics: Huang sees AI evolving into "agentic AI," where AI systems can perceive, reason, plan, and act autonomously. He believes this represents a "multi-trillion-dollar opportunity" for NVIDIA, transforming industries with AI agents acting as a digital workforce
  • ​Physical AI: Huang envisions the next stage of AI development as "physical AI," which involves applying AI to the tangible world, such as robots and fully autonomous vehicles. This phase is expected to revolutionize robotics and autonomous systems
  • NVIDIA's Platforms for Robotics: NVIDIA has introduced platforms like Cosmos, Omniverse, and Isaac GR00T to support AI development in robotics. These platforms simulate real-world environments and facilitate robot learning and development, positioning NVIDIA for dominance in the robotics market.
  • NVIDIA reported a record quarterly revenue of $39.3 billion for the fourth quarter of fiscal 2025, marking a 12% increase from the previous quarter and a 78% rise from the same period last year.

AI Use Cases in Business & Insurance

• Lyft is integrating Anthropic’s Claude-powered AI to enhance automation, reducing customer service resolution times. This collaboration has already reduced customer service resolution times by 87%.
• Generali is partnering with MIT's Laboratory for Information and Decision Systems, focusing on AI innovations in risk modeling, claims assessment, and underwriting
• Moody's is partnering with hyperexponential to improve pricing and underwriting for reinsurers using advanced analytics via hx Renew

AI Research & Development

• Neuralink’s patient successfully controlled a robotic arm using only their thoughts.
• Anthropic introduced Claude 3.7 Sonnet, combining large language model (LLM) capabilities with reasoning algorithms in a single system (The Deep View, 25 Feb 2025).
• Stanford researchers released Evo 2, an open-source AI model trained on 9 trillion nucleotides that can generate new genetic code (The Deep View, 24 Feb 2025).
• Google launched an AI-powered scientific research assistant, built on Gemini 2.0, which can generate and validate scientific hypotheses (The Rundown AI, 20 Feb 2025).
• Figure unveiled a new AI Vision-Language-Action model, Helix, which enables robots to understand speech and handle objects they've never seen before (The Rundown AI, 21 Feb 2025).
• Neuralink’s patient successfully controlled a robotic arm using only their thoughts (Lore, 06 Feb 2025).

AI Regulation & Ethics

• Switzerland has decided to ratify the Council of Europe's AI Convention and implement sector-specific regulations, focusing on areas like data protection and transparency. There is no immediate plan for a "Swiss AI Act" mirroring the EU's framework, but Switzerland aims to ensure compliance with EU regulations for market access.
• Europe might tighten regulations on AI-generated content transparency (simple.ai, 12 Feb 2025).
• EU dropped the AI Liability Directive. The withdrawal reflects ongoing challenges in establishing clear AI liability frameworks. It leaves the EU without a specific directive addressing AI-related damages, relying instead on broader regulations like the revised Product Liability Directive (Sifted Newsletter, 13 Feb 2025).
• South Korea halted DeepSeek chatbot due to regulatory concerns (Eye on AI | Fortune, 18 Feb 2025).
• The AI Action Summit in Paris ended with rising global tensions as the U.S. and U.K. refused to sign a multinational declaration on open, ethical AI (Eye on AI, 11 Feb 2025, The Rundown AI, 12 Feb 2025).

​

​Unsubscribe · Preferences​

​