Doublebrass's question on Reward Hacking

From Stampy's Wiki
Doublebrass's question on Reward Hacking id:UgzmAXIFVZqOq5 ZzZ4AaABAg /
Revision as of 19:40, 26 April 2021 by 756254556811165756 (talk | contribs) (fixing dates with regex)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)

Super interesting! If this kind of reward hacking exists in current AI, does that have any kind of serious implications if someone wanted to deploy one for the stock market, for example? Like would the AI seek to "cheat" and commit fraud or some gain insider info rather than play the stock market fairly?

Tags: None (add tags)
Question Info
Asked by: doublebrass
OriginWhere was this question originally asked
YouTube (comment link)
On video: Reward Hacking: Concrete Problems in AI Safety Part 3
Date: 2017-08-12T19:40
Asked on Discord? No
YouTube Likes: 2
Reply count: 4