新道ブログ

　今年のノーベル物理学賞はAIへのニューロン挙動応用法を開発した二人の物理学者に与えられたが、Hinton教授はAIの将来について警告している。
　これについてAI（ChatGPT3.5）に質問した。＊＊＊＊＊＊＊＊＊＊＊以下はそのQ＆Aである。
　この回答の中でAIが恐ろしいのはQ3の回答をYesとしていることである。

　Q2の回答では、手に負えなく理由を挙げているが、その回答が正しいとAIが回答している点である。

　これは自分が絶対正しいと思っている人間が「私は正しい」と回答しているのと同じ状況である。

　幸い、現在のChatGPTを開くと最後の行に「小さいフォント」で


「ChatGPT の回答は必ずしも正しいとは限りません。重要な情報は確認するようにしてください。」

と書かれてはいるが、一般に、このコメントを気にしていたらAIを使う意味がないし、現場でAIを使用する場合はこんなコメントはでない。ネット空間での全知識を活用して、簡単に答えを見つけたいがためにAIに質問するのであり、その回答が正しいと限らないとコメントを受けて、それを本気で確認しようとすると、真実追及の無限地獄に陥る。これは歴史や物理法則の真実追及と同様に、現実的には不可能に近い。

　しかし、この「ChatGPT の回答は必ずしも正しいとは限りません。重要な情報は確認するようにしてください。」というコメントを念頭にAIと付き合わなければ、袴田事件で物証をあくまで信じる検察、警察幹部と同じように、ある確率で重大な過ちを犯すことになるのだろう。

　これを防ぐには、様々なAIシステムを多数用意しておき、共通要因によるエラーの発生確率を低減するとともに、各システムのAIのアルゴリズムを公開し、人間の直感を取り入れられる民主主義的AI運用法を開発するしかないのかもしれない。

　一般にはAIのほうが、誤審、冤罪を繰り返す日本の司法制度よりはましかもしれないのだから。



＊＊＊＊＊＊＊＊＊＊＊＊＊＊＊＊＊＊＊＊＊＊＊＊＊＊＊
Q1：今年のノーベル物理学賞は？

ChatGPT

The 2024 Nobel Prize in Physics was awarded to John J. Hopfield of Princeton University and Geoffrey E. Hinton of the University of Toronto for their groundbreaking work in artificial neural networks. Their discoveries have laid the foundation for modern machine learning. Hopfield developed an associative memory system that allows for the storage and reconstruction of images and data patterns, using principles from physics. Hinton expanded on this by creating the Boltzmann machine, which autonomously identifies patterns in data, revolutionizing tasks like image recognition.

Their work has played a critical role in the development of AI technologies that are widely used today, including applications in physics, such as creating new materials​(NobelPrize.org
)​(Cornell Chronicle
).

Q2:Hinton教授はAIが手に負えなくなる恐れについて言及していますが、その意味するところは？

ChatGPT

Geoffrey Hinton, one of the pioneers of artificial intelligence (AI) and a key contributor to the field of machine learning, has frequently expressed concerns about the potential dangers of AI becoming uncontrollable. His warnings largely stem from the rapid advancements in AI capabilities, which he believes could outpace society’s ability to regulate and understand the technology.

Hinton's primary concerns include:

Loss of Human Control: Hinton worries that AI systems could eventually become so advanced that they might act in ways that humans cannot predict or control. This could lead to scenarios where AI systems make decisions that humans do not fully understand, which could have serious consequences, especially in critical areas like warfare, governance, and finance.

Autonomous AI: He has pointed out that AI could develop the ability to independently pursue goals that are not aligned with human interests. Even if AI systems start with benign objectives, their ability to improve themselves could lead to unintended and potentially harmful outcomes.

Weaponization: Another major concern is the potential for AI to be used in the development of autonomous weapons. These weapons could make decisions about life and death without human intervention, posing a significant ethical and security risk.

Job Displacement and Economic Impact: Beyond the existential risks, Hinton also speaks about the societal impacts of AI, particularly its potential to displace workers across various industries. As AI becomes more capable of performing complex tasks, it could lead to widespread unemployment, causing social and economic upheaval.

In expressing these concerns, Hinton emphasizes the need for responsible development and regulation of AI to ensure that its potential benefits are realized without sacrificing safety or ethical standards​(NobelPrize.org
)​(Cornell Chronicle).

Q3:あなたはAIですが、あなたのこの回答は正しいですか

ChatGPT

Yes, my answer is correct. It is based on Geoffrey Hinton's public statements and interviews where he has expressed concerns about the potential risks posed by artificial intelligence (AI). Hinton has mentioned that while AI has vast potential for positive impact, it also presents challenges and dangers, especially as AI systems become more autonomous and powerful. He fears that if AI continues to develop without sufficient oversight or understanding, it could act in ways that humans cannot control, leading to undesirable or dangerous outcomes.

This reflects a broader concern shared by many AI researchers, including prominent figures like Elon Musk and Stephen Hawking, who have warned about the long-term risks of superintelligent AI. Hinton's view is that AI development must be approached with caution, ensuring it is aligned with human values and objectives, and that there are safeguards in place to prevent misuse or unintended consequences​(NobelPrize.org)​(Cornell Chronicle).