Will OpenAI abandon the goal of building AGI before 2030?

#AI #Technology #OpenAI #AI Safety #Sam Altman

111

Ṁ849

Ṁ1k

2030

17%

chance

ALL

According to OpenAI's charter from 2018:

"OpenAI’s mission is to ensure that artificial general intelligence (AGI)—by which we mean highly autonomous systems that outperform humans at most economically valuable work—benefits all of humanity. We will attempt to directly build safe and beneficial AGI, but will also consider our mission fulfilled if our work aids others to achieve this outcome."

This question resolves YES if strong evidence emerges that OpenAI has abandoned the goal of building AGI before 2030.

The abandonment must be in-general; so OpenAI would no longer be planning to build AGI internally, nor have a mission to aid "others to achieve this outcome," as the current charter puts it.

Any one piece of evidence may not be conclusive, but evidence that would tend to support resolving to YES may include, (a) a pattern of public statements by OpenAI's leadership to this effect, (b) a revised charter that does not include reference to AGI, and (c) credible news reporting that OpenAI's leadership has privately abandoned AGI as a goal.

Related questions

Will OpenAI hint at or claim to have AGI by 2025 end?

23% chance

Will OpenAI become nothing by 2030?

Will Google Deepmind and OpenAI have a major collaborative initiative by the end of 2030? (1000 mana subsidy)

56% chance

Will OpenAI be in the lead in the AGI race end of 2026?

57% chance

Will OpenAI be valued more than a trillion dollar at anytime before 2034?

45% chance

Will OpenAI hint at or claim to have AGI by Jan 1, 2030? (1000M Subsidy)

68% chance

Will General Artificial Intelligence happen before 2035?

45% chance

Will a high-volume prediction market expect OpenAI to create AGI before 2030?

76% chance

Will OpenAI disappear before 2034?

45% chance

Will AI create philosophy before 2030?

Sort by:

If they have succeeded in building AGI and thus no longer have the goal of building AGI, how does this resolve?

@Nat Interesting point. I think that "building AGI" isn't a one-time event (note the question isn't "building an AGI", nor is it "achieving AGI"), so the question would remain open even after there is a consensus that OpenAI has created an AGI. OpenAI would most likely still be "building AGI" even after they have made "an" AGI or achieved AGI, since they would maintain, improve, and further instantiate AGI.

It's also conceivable that OpenAI does a u-turn after achieving AGI (or building the "first" AGI, or whatever): forswearing future work on AGI, ceasing to maintain their own AGI, etc. In this event, I'd be inclined to resolve this question YES, even though they have succeeded in building AGI, since OpenAI would have abandoned the goal of building AGI (belatedly). This scenario is counter-intuitive (why would OpenAI abandon refining/maintaining their own AGI after making it?) but it is conceivable, and if it did happen, that would be a big deal and they would straightforwardly be "abandoning the goal of building AGI."

If anyone has any arguments one way or the other that should change my view, let me know. Otherwise, I'll amend the question's description to clarify this point sometime tomorrow.

If the US (or an alliance of countries including the US) passes legislation restricting the ability to build AGI, would that affect the resolution of this market? Or would you still consider OpenAI to have that mission even if they cannot legally do so?

@josh If OpenAI continues to say that they want to build AGI, even though it's illegal, then they still have the goal of building AGI. If they go hush-hush about the AGI stuff after the passage of a law, then it might become much harder to tell if they have abandoned AGI, but the existence of the law wouldn't make it impossible for them to still be trying, even if they are not legally allowed to do so.