Statement on AI Risk

Re:Statement on AI Risk by blueblimp on Tuesday May 30, 2023 @08:32AM (#36124039)

There is a way, in my opinion: distribute AI widely and give it a diversity of values, so that any one AI attempting takeover (or being misused) is opposed by the others. This is best achieved by having both open source and a competitive market of many companies with their own proprietary models.

[ Parent | on HN ]

Re:Statement on AI Risk by ChatGTP on Tuesday May 30, 2023 @09:28AM (#36124640)

How do you give "AI" a diversity of values?

[ Parent | on HN ]

Re:Statement on AI Risk by drvdevd on Tuesday May 30, 2023 @02:05PM (#36128264)

By driving down the costs of training and inference, and then encouraging experiments. For LLMs, QLoRA is arguably a great step in this direction.

[ Parent | on HN ]

Re:Statement on AI Risk by blueblimp on Tuesday May 30, 2023 @06:36PM (#36132015)

Personalization, customization, etc.: by aligning AI systems to many users, we benefit from the already-existing diversity of values among different people. This could be achieved via open source or proprietary means; the important thing is that the system works for the user and not for whichever company made it.

[ Parent | on HN ]

Re:Statement on AI Risk by q1w2 on Tuesday May 30, 2023 @09:48PM (#36133573)

Like giving nuclear weapons to everyone. What could go wrong?

[ Parent | on HN ]

Re:Statement on AI Risk by blueblimp on Wednesday May 31, 2023 @01:44AM (#36134899)

AI isn't nuclear weaponry. It's bits, not atoms. It's more like encryption.

[ Parent | on HN ]