Next phase of AI tech boom // not AGI, but what?
How do you prove that a LLM is actually correct and not hallucinating?
At the World Economic Forum in Davos, Switzerland, generative artificial intelligence (AI) takes center stage as major tech companies like Salesforce, Microsoft, and Google showcase their advancements. While 2023 showcased AI’s capabilities, 2024 is expected to focus on improving accuracy, particularly in high-stakes industries like healthcare and manufacturing. Intel CEO Pat Gelsinger emphasizes the need to build formal correctness into AI models for applications such as medical diagnoses, manufacturing quality control, and automated driving. The key challenge lies in proving the accuracy of large language models. Salesforce AI CEO Clara Shih suggests a three-phase approach to AI adoption, involving active use, monitoring, and eventually trusting the technology. Open AI CEO Sam Altman views AI as a productivity tool rather than a job replacement. Nasdaq CEO Adena Friedman highlights AI’s potential in updating old code and modernizing systems in the financial industry. The overall sentiment at Davos is optimistic, with a focus on experimenting, co-piloting, and gradually trusting AI for enhanced productivity and efficiency.
