This is a special post for quick takes by Avani Gupta. Only they can create top-level comments. Comments here also appear on the Quick Takes page and All Posts page.
Is cooperation in AI agents desirable? Cooperation when geared towards human betterment/intentions - arguably desired* Cooperation when against our purposes (serving something else like cheating for better grader output) is un-desired and we call it collusion.
*Imagine a world where agents act on your behalf and don't cooperate with each other acting in selfish ways. Wouldn't the societies collapse? Human world possibly sustains on cooperation in a lot of ways - we cooperate to use and replenish resources and maintain harmony in different ways. Human organizations and corporations are in some way an effort towards maintaining the collective harmony. Is'nt the AI safety effort (or existence of LessWrong) a consequence human cooperation to prevent AI risks?