最后，智能体还需要有很强的可靠性、可控性，才能规模化的带来价值。这既体现在智能体能否稳定、鲁棒地完成任务，也体现在其行为是否始终与人类的真实意图与价值观保持一致。在《2001：太空漫游》中，AI为完成任务选择牺牲人类乘员，正是目标函数与人类价值未能有效对齐的极端后果。随着智能体智能水平与自主性的提升，这类对齐失败带来的风险可能会被进一步放大。
Subscribe to get our editors’ pick of the Guardian’s award-winning sport coverage. We’ll email you the stand-out features and interviews, insightful analysis and highlights from the archive, plus films, podcasts, galleries and more – all arriving in your inbox at every Friday lunchtime. And we’ll set you up for the weekend and let you know our live coverage plans so you’ll be ahead of the game. Here’s what you can expect from us.
。safew官方下载对此有专业解读
В Финляндии предупредили об опасном шаге ЕС против России09:28，详情可参考同城约会
（一）结伙斗殴或者随意殴打他人的；，详情可参考同城约会