Member
- Joined
- Oct 14, 2023
- Messages
- 225
- Thread Author
- #1
Creating a perfect AI that looks like a human can negatively affect the future of people.
In a new documentary released by The Guardian, Ilya Sutskever, co-founder and chief scientific officer of OpenAI, expressed the opinion that the technology developed by his company will change the world, but not always in favor of humanity. Shot between 2016 and 2019, the film shows Sutskever during the period when OpenAI was creating the technologies behind the existing popular ChatGPT. He notes that artificial intelligence (AI) will solve many of today's problems, including employment, disease, and poverty, but it will also create new ones.
.
Sutskever, known for his machine learning research, said in early 2022 that some large language models (LLMs) may be conscious. He emphasizes the importance of "AI alignment" – the task of creating an AI that will meet our goals.
Commenting on the possible goals of AI, Sutskever compares them to the behavior of humans in relation to animals – we do not ask permission from animals when we build highways, because it is convenient for us. According to Sutskever, this is exactly the relationship that can develop between people and Artificial General Intelligence (AGI) if it is not programmed correctly.
Sutskever warns that if AI programming is done incorrectly, the nature of evolution and natural selection will help those systems that prioritize their own survival survive. The expert concludes that this is extremely important for the future of humanity.
In a new documentary released by The Guardian, Ilya Sutskever, co-founder and chief scientific officer of OpenAI, expressed the opinion that the technology developed by his company will change the world, but not always in favor of humanity. Shot between 2016 and 2019, the film shows Sutskever during the period when OpenAI was creating the technologies behind the existing popular ChatGPT. He notes that artificial intelligence (AI) will solve many of today's problems, including employment, disease, and poverty, but it will also create new ones.
Sutskever, known for his machine learning research, said in early 2022 that some large language models (LLMs) may be conscious. He emphasizes the importance of "AI alignment" – the task of creating an AI that will meet our goals.
Commenting on the possible goals of AI, Sutskever compares them to the behavior of humans in relation to animals – we do not ask permission from animals when we build highways, because it is convenient for us. According to Sutskever, this is exactly the relationship that can develop between people and Artificial General Intelligence (AGI) if it is not programmed correctly.
Sutskever warns that if AI programming is done incorrectly, the nature of evolution and natural selection will help those systems that prioritize their own survival survive. The expert concludes that this is extremely important for the future of humanity.