IEEE Spectrum compiled a scorecard summarizing the positions of 22 AI authorities on large language modelsthe likelihood of AGI, and the risk of civilizational catastrophe. 8 think LLMs can lead to AGI, and 4 think AGI could lead to doomsday.
AI experts remain divided over the implications of large-scale language models, such as OpenAI’s GPT-4, for the creation of artificial general intelligence (AGI) and the potential risks they pose. Some experts worry about the disruptive effects of generative AI on employment, bias, and the spread of misinformation, while others dismiss apocalyptic scenarios as science fiction.