[pipthepixie]

> We want to take this work a step further by inviting and incentivizing the community to help identify potential harms of this algorithm beyond what we identified ourselves.

What more needs to be done here? Twitter got caught out with bias in their image cropping algorithm, and now they want people to further elaborate on how/why this is?

> For this challenge, we are re-sharing our saliency model and the code used to generate a crop of an image given a predicted maximally salient point and asking participants to build their own assessment.

Seems like a very exhaustive endeavour just to point out algorithmic bias and re-iterate what went wrong, but in more detail.

[SpicyLemonZest]

They want to develop a fuller understanding of what kinds of biases they need to be aware of. "Caught out with bias" isn't a good way to understand the situation - this isn't the kind of problem that can be solved by just debugging to find the biased functions and using unbiased ones instead.

[vineyardmike]

> Seems like a very exhaustive endeavour just to point out algorithmic bias and re-iterate what went wrong, but in more detail.

They want everyone to help them find additional bias, and other points it could go wrong, with the presumed purpose of them fixing it.

They want to find bias they can solve and remove, not just get more people to tell them why the bias they found is wrong!