ChatGPT reached 100 million users in January 2023, making it the fastest-growing consumer app in history. GPT-4 is rumored to have around 1.76 trillion parameters across eight Mixture-of-Experts models. Claude 2 offers up to 100,000 tokens of context and uses Constitutional AI to encourage harmless, helpful behavior. Gemini Ultra reportedly exceeded 90% on the MMLU…
Read more
The 2017 U.S. Department of Defense project Maven was launched to apply AI-based computer vision to analyze drone surveillance footage in the Middle East, with fielding by year’s end. In 2019, China released its National Defense White Paper announcing the goal of “intelligentized warfare” and a plan to become the world AI leader by 2030.…
Read more
Executive snapshot 1. Where the law stands today (mid-2025) Jurisdiction Binding instrument Status & key 2025 deadlines Direct requirements for generative / foundation models European Union Regulation (EU) 2024/1689 – “AI Act” Entered into force 1 Aug 2024; prohibitions apply 2 Feb 2025; general-purpose model rules kick in 2 Aug 2025 and most other obligations…
Read more
The European Union finalized the AI Act in 2024, with full applicability by 2026, classifying AI by risk and requiring high-risk systems to include transparency, human oversight, and clear labeling for AI-generated content including deepfakes. China enacted the Interim Measures for Generative AI Services in August 2023, requiring provider registration, alignment with core socialist values,…
Read more
Over 70% of large firms plan to invest in AI-driven security tools by 2027. In early 2022, a multinational technology manufacturer faced Babuk ransomware and an AI-driven defense autonomously blocked and isolated the infected device. A healthcare company using Darktrace’s UEBA detected an insider attempting data theft by connecting an employee’s device to the Dark…
Read more
OpenAI launched GPT-4.5 “Orion” on February 27, 2025, as the largest model to date with improved reasoning and a chain-of-thought mode. Google’s Gemini 2.5, released in 2025, is a multimodal AI that handles text, code, images, audio, and video and comes in Nano, Flash, Pro, and Ultra tiers with an AI Pro plan and Deep…
Read more
In May 1999, NASA’s Deep Space 1 operated for three days with the Remote Agent AI, planning activities and diagnosing simulated faults autonomously. From 2001 to 2004, NASA’s EO-1 carried the Autonomous Sciencecraft Experiment (ASE), using onboard machine learning and the CASPER planner to re-task after events like volcanic eruptions. In 2013, JAXA’s Epsilon rocket…
Read more
