AI in October: Powering Up & Cracking Down
Big Tech Turns to Nuclear for AI’s Power Demands
The demand for clean, reliable energy to power artificial intelligence is prompting major tech companies to invest in nuclear solutions. Google recently made headlines by partnering with Kairos Power to deploy small nuclear reactors by 2030, providing a sustainable energy source for its data centers. Michael Terrell, Google’s Senior Director for Energy and Climate, emphasized that nuclear will help meet AI’s “clean and reliable” energy needs, reflecting the company’s commitment to low-carbon energy solutions as AI-driven power demands increase.
Amazon Web Services (AWS) is also investing $500 million in small modular reactors (SMRs) through partnerships with Dominion Energy and Energy Northwest, making it the third major tech company, after Google and Microsoft, to focus on nuclear energy. AWS CEO Matthew Garman highlighted the necessity of nuclear, pointing out that renewable sources alone may not meet the energy needs of AI, which requires constant, stable power. These investments show a significant pivot towards nuclear as the tech industry prepares for an AI-driven future that demands both sustainability and reliable power sources.
Microsoft and Anthropic Elevate AI Capabilities
Microsoft’s recent updates to its Copilot assistant and Anthropic’s new Claude 3.5 models mark significant advancements in AI’s capabilities. Microsoft reimagined its Copilot to include voice and vision functions, aiming to make it a more interactive, conversational assistant. Users can now interact with Copilot through a “Copilot Daily” feature for audio news updates and ask questions about on-screen content, with the AI adapting naturally to user needs. This update is part of CEO Mustafa Suleyman's vision to make Copilot a warmer, more integrated experience for users.
Leading into these developments, Anthropic introduced its upgraded Claude 3.5 Sonnet and Haiku models, offering groundbreaking features like “computer control” that allow the AI to navigate screens, click, and type just as a human would. The Sonnet model, which excels in coding, now sets the stage for AI to perform complex tasks with higher efficiency and precision. Meanwhile, Claude 3.5 Haiku focuses on affordability and speed without compromising safety, underscoring Anthropic’s commitment to responsible scaling. Together, Microsoft and Anthropic are pushing the boundaries of what AI can achieve in everyday and professional environments.
Government’s Balancing Act with AI Regulation and Fraud Prevention
California Governor Gavin Newsom recently vetoed a major AI safety bill, stalling some of the first proposed AI regulations in the U.S. Authored by Senator Scott Wiener, the bill would have introduced requirements for safety testing and included a “kill switch” to shut down potentially dangerous systems. Newsom, however, argued that the bill’s broad standards could harm innovation and deter AI companies from operating in California. While he dismissed this specific legislation, Newsom signaled plans to work on targeted AI safeguards, engaging experts to develop a balanced approach.
Meanwhile, the U.S. Treasury has successfully leveraged AI to combat fraud, reclaiming $4 billion in improper payments and fraud in 2024—significantly up from $652.7 million the previous year. This effort is part of the Treasury’s ongoing work to protect taxpayer dollars, with AI tools now able to detect complex patterns that were previously hard to identify.
In the political realm, a Clemson University study recently found that AI-powered bot networks have been spreading pro-Republican messages on X (formerly Twitter). Using language models like ChatGPT, the bots targeted key states and races, amplifying political narratives. While there is no evidence of foreign interference, the AI-driven activity highlights the ethical challenges governments face as AI becomes more embedded in both political discourse and security operations.
Sources:
BBC News: "Newsom vetoes AI regulation bill amid industry pushback"
The Verge: "Microsoft redefines Copilot with voice and vision"
Cointelegraph: "AI helps Treasury recover billions in fraud"
Artificial Intelligence News: "Anthropic’s new Claude AI models and computer control feature"
NBC News: "Amazon’s nuclear investment for AWS data centers"
