Generative AI & LLMs
Model welfare
Discussion about whether AI could have experiences or needs that people should consider ethically.
Example
A research program investigates possible indicators of AI system welfare.
Why people use it
It investigates whether AI might have experiences that deserve ethical consideration.
What you'll hear
“What evidence would suggest an AI has welfare interests?”
What this means for you
Distinguish precautionary research from a claim that sentience has been established.
Can you control it?
No
No direct control. This describes a wider issue, concept or result rather than something you can simply switch on or off in a tool.
Common questions
- Does studying AI system welfare establish that current AI systems can suffer?
- No. The subject explicitly involves substantial uncertainty.
- Does a model requesting protection settle the issue?
- No. Generated requests do not by themselves establish feelings or interests.
- Can people investigate without agreeing on the answer?
- Yes. Research can examine possible indicators while acknowledging substantial uncertainty.