Helen Toner, a notable figure in AI, recently experienced a tumultuous tenure at OpenAI, marked by the controversial dismissal of CEO Sam Altman. Now at Georgetown University's CSET, she critiques the prevalent misconception of AI as infallible, tracing it back to science fiction ideals. In her perspective, the general perception of AI's capabilities falls short of reality due to its non-deterministic nature, which contrasts sharply with our experiences with traditional computing systems. Toner remains engaged in discussions about the future of AI, particularly in the defense sector.
I would say there are two reasons for the 'magical' properties ascribed to AI. One on us is learned through science fiction, where AI is usually infallible.
Toner emphasizes the distinction between traditional computers and AI, noting that while calculators are reliable, AI systems are non-deterministic and only provide pattern-based outputs.
Collection
[
|
...
]