Ilya Sutskever warns that advanced AI reasoning will lead to unpredictability

VANCOUVER, Dec 13 (Reuters) - Ilya Sutskever, former chief scientist at OpenAI, stated that AI systems with reasoning capabilities will become less predictable. He spoke at the NeurIPS conference, noting that the current method of pre-training AI is reaching its limits. Sutskever highlighted that while computing power is increasing, the amount of available data is not. He suggested that future AI could generate new data or evaluate multiple answers to improve accuracy. He predicted that advanced AI will reason like humans, but this reasoning will lead to unpredictability. Sutskever cited examples from games like Go and chess, where AI has surprised experts with unexpected moves.