埃文锋利

2023年3月29日,

Elon Musk,史蒂夫·沃兹尼亚克,Pinterest创始人埃文·夏普,稳定AI CEO Emad Mostaque都增加了他们的签名发表一封公开信的未来生活研究所,一个非营利性工作从强大的技术降低存在风险。信中警告说,人工智能系统,如OpenAI GPT-4正在成为“人类极少在通用任务”,对人类和社会构成潜在风险。它要求AI实验室暂停训练比GPT-4技术更强大了六个月,正确评估新技术的危险。业内专家Yoshua Bengio,有时被称为一个“人工智能的教父,”和影响力的计算机科学家斯图尔特·罗素也把他们的支持信。发表的时候,不代表OpenAI似乎已经签署了这封信。信中引用担心错误信息的传播,自动化在劳动力市场上的风险,和文明的失控的可能性。要点:失控的AI非营利浮动的可能性开发商失去控制强大的文明新的人工智能系统和他们的预期效果。这表明公司正竞相开发人工智能技术如此先进,以至于即使是创作者可以“理解、预测或可靠的控制。”The letter stated: "Should we risk loss of control of our civilization? Such decisions must not be delegated to unelected tech leaders." A "dangerous race" The letter warned that AI companies are locked in an "out-of-control race to develop and deploy" new advanced systems. In recent months, the viral popularity of OpenAI's ChatGPT has appeared to push other companies to release their own AI products. The letter urged companies to reap the rewards of an "AI summer" while society has a chance to adapt to the new technology instead of rushing into an "unprepared fall." AI automation and misinformation The letter highlighted several risks of the new tech, including the possibility that nonhuman minds will eventually "outnumber, outsmart, obsolete and replace us." It said that AI systems are becoming "human-competitive" at some tasks and cited concerns around misinformation and labor automation, stating: "Should we let machines flood our information channels with propaganda and untruth? Should we automate away all the jobs, including the fulfilling ones?" Six-month pause The open letter asks for a six-month break from developing any AI systems more powerful than those already on the market. It asks developers to work with policymakers to create AI governance systems, highlighting the need for regulatory authorities as well as AI "watermarking systems" to help people differentiate between human and AI-made content. The letter also suggests the need for "well-resourced institutions" to cope with economic and political disruptions caused by AI. The open letter stated the pause should be a step back from a "dangerous race" around advanced technology rather than a complete stop on general AI development. Sign up for notifications from Insider! Stay up to date with what you want to know. Subscribe to push notifications