These are the kinds of spots where software tools can help us a lot.
To be clear, if we weren't on the stone bubble, with another pretty short stack also in play, this would be a very clear call off. But since we're on the bubble I'm gonna look at ICMizer since we don't always have perfect intuition for how the bubble should affect our play.
My instinct is that AJo would be in the lower middle or upper bottom part of a chipEV call off range. And I think there is fairly high ICM pressure on us due to there being multiple shorties. I would LEAN toward a fold but this would be the very best hand I fold. Now let's look at ICMizer.
https://www.icmpoker.com/icmizer/#vqHoDV
Well it's a bit misleading because ICMizer thinks that the HJ should jam 24% of hands, and there's just no way a guy with those stats is jamming that wide here, but what's interesting is that even with the villain jamming THAT wide, we're still only supposed to call with 11% of hands due to ICM (vs 17% chipEV calling range).
What happens when we change the villain's range to a more realistic 15%? We are still supposed to call with AJo interestingly, but it's right on the edge.
https://www.icmpoker.com/icmizer/#vRJsMl
But it's the bottom effectively. And is this guy even jamming 15%? He has opened 4% of hands in a 120 hand sample AND it's the bubble and there are shorter stacks than his but he's covered by BOTH blinds. I think it's more likely he jams something like 9% (66+, AJ+, KQ). In which case we need to fold AJ, easily, and AQo is a fold but on the borderline.
https://www.icmpoker.com/icmizer/#vIIIJx
I think the last scenario is the closest to realistic, so I think AJ is a fold here. But it is at least close depending on how accurate you think your stats are on the guy and how well they apply to the bubble. I would think he's on the tighter end of this spectrum and I'd be sigh folding AJ here typically.