在3月24日于北京钓鱼台国宾馆隆重召开的中国发展高层论坛上，图灵奖得主约瑟夫·希发基思就人工智能的发展与治理发表了深入见解。希发基思，一位在计算机科学领域享有盛誉的专家，指出尽管人工智能（AI）研究已经取得了显著的突破，特别是在生成式AI方面，但目前实际应用的技术仍然停留在弱人工智能（weak AI）阶段。
希发基思在演讲中强调，当前的AI系统主要集中在问答等特定应用上，它们能够处理和回答复杂的问题，但这并不意味着它们具备了全面的智能。他表示，现有的AI并未发展出与人类同步的智能能力，即理解、学习和适应新环境的能力。他提醒，尽管AI在某些领域展现出惊人的潜力，但我们尚未掌握能够复制人类全面智能的技术和法则。
这一观点在论坛上引发了深入的讨论，与会者就如何在推动AI技术进步的同时，确保其治理和应用的合理性与安全性展开了广泛对话。希发基思的发言提醒人们，尽管AI的进步值得庆祝，但对其能力的评估和未来的研发方向应保持审慎和务实的态度。
英语如下：
**News Title:** “Turing Award Winner Warns: Current AI Remains at Narrow Intelligence, Generative Applications Limited to Question-Answering”
**Keywords:** Joseph Sifakis, Narrow Artificial Intelligence, Turing Award
**News Content:**
**Title:** Turing Award Winner Joseph Sifakis: Current AI Remains in the Realm of Narrow Intelligence
At the China Development Forum held on March 24 at the Diaoyutai State Guesthouse in Beijing, Joseph Sifakis, a Turing Award recipient and esteemed computer science expert, shared his insights on the development and governance of artificial intelligence (AI). Sifakis pointed out that while AI research has seen remarkable breakthroughs, particularly in generative AI, practical applications still dwell in the realm of narrow AI (weak AI).
In his speech, Sifakis emphasized that current AI systems primarily focus on specific applications like question-answering, capable of addressing complex inquiries. However, he cautioned that this does not equate to comprehensive intelligence. He explained that existing AI has not yet developed the human-like capabilities of understanding, learning, and adapting to new environments. He reminded the audience that while AI demonstrates astonishing potential in certain domains, we have yet to grasp the technologies and principles to replicate human-like general intelligence.
This perspective sparked in-depth discussions at the forum, with participants engaging in wide-ranging conversations about advancing AI technology while ensuring its responsible and secure application. Sifakis’ remarks underscore the importance of maintaining a cautious and pragmatic approach in assessing AI’s capabilities and guiding future research directions.
【来源】https://economy.caixin.com/2024-03-24/102178947.html
Views: 1