And it’s time for our last Friday Big Ideas of the year.
Starting with Schneier on Security in an article called AI and Trust. Trust is essential to society, and there are two kinds of trust: interpersonal and social. The confusion between the two will increase with artificial intelligence (AI), as we will think of AIs as friends when they are actually just services. The corporations controlling AI systems will take advantage of this confusion, making it important for the government to create an environment for trustworthy AI through regulation. The market alone cannot provide trustworthy AI, as corporations prioritize profit. AI transparency laws, regulations on AI safety and security, and the concept of data fiduciaries are necessary. Additionally, public AI models built by the public for the public can serve as a counter-balance to corporate-owned AI. Government plays a crucial role in creating social trust and enforcing predictability and reliability.
I wanted to point you to an article in Ars Technica that discusses the research behind OpenAI’s Q* project, which aims to improve the reasoning capabilities of large language models. It explores techniques such as chain-of-thought prompting, step-by-step verification, and tree search. The article also speculates on the future direction of this research and the challenges that need to be overcome. It’s a guided tour of the research.
Finally, from Dark Reading, Sanjay Beri offers commentary on implementing AI. Let me first quote.
One leader at a large financial firm recently told me the board is very eager to leverage generative AI: “It’s a competitive advantage. It’s the key to automation. We have to start using it.” But when I asked what they’re doing with AI, they replied, “Oh, we’re blocking it.”
If this sounds like you, Beri discusses the responsible implementation of AI, highlighting the positive and negative aspects. The article emphasizes the importance of securely enabling AI and controlling sensitive data. It suggests four key points for moving forward, including accepting the inevitability of AI, focusing on responsible usage, ensuring efficacy and accuracy, and classifying the sensitivity of AI applications.
Why do we care?
This should be excellent reading to get you thinking this holiday.