Seems worthwhile and good for an extra reason: It allows us to train AIs on only data from before there was significant AI generated content, which might mitigate some safety concerns around AIs influencing future AIs training.
For anyone seeking more information on this, feel free to search up the key terms ‘data poisoning’ and ‘Trojans.’ The Centre for AI Safety has somewhat accessible content and notes on this under lecture 13 here.
Seems worthwhile and good for an extra reason: It allows us to train AIs on only data from before there was significant AI generated content, which might mitigate some safety concerns around AIs influencing future AIs training.
For anyone seeking more information on this, feel free to search up the key terms ‘data poisoning’ and ‘Trojans.’ The Centre for AI Safety has somewhat accessible content and notes on this under lecture 13 here.