This is the worst way forward with AI. It is unethical to fund or publish scholarship like this. https://twitter.com/forcriticaltech/status/1275402019130478595

#defundAI if it’s gonna be like this (& it’s already shown us over & over again that it is)

People aren’t struggling to #DefundThePolice just to lose their rights to AI. It’s beyond fucked up how the more we discuss the need for ethical, humane AI the more we see shit like this

It looks like springer has now backed away from publishing this paper due to uproar—which is good BUT this project isn’t unique. This keeps happening. These things keep getting through peer review. If peer review is this broken in AI then funding cuts are the only way to stop it.

This happens when a field is overfunded at institutional & external grant levels—the scrutiny & rigor just aren’t there. Everyone w/half an idea, or terrible ideas, gets funded, handsomely. And it’s a constant game of catch up to reduce the harms they’ve been paid to cause...

the level of salary and grant funding should not so vastly exceed a field’s ability to produce cogent work. In the humanities so many amazing projects go unfinished/undone because there is NO funding. Something like 10-15% of projects get funded, w/usually tiny grants of $2-10k

That also hurts the rigor of fields, because we lose so many good projects that never get done—disproportionately ones by marginalized scholars. Meanwhile fields firmly aligned with militarism, surveillance, & policing swim in money, doing vast harm through substandard research.

so when I say #defundAI I’m not joking. The researchers are failing. Their peer reviewers are failing. Their publishers & employers are failing. We can’t play catch-up to undo harms a privileged few are creating, designing bad tech. We need to cut off the support. Now’s the time.

Again, this has been going on forever, so it’s not like it just started & it’s a simple problem to fix. It’s not a bug in the system, it is the system. Read @merbroussard’s award winning book Artificial Unintelligence for a history of how AI funding was problematic from the jump.

I spend so much time in the classroom teaching history & STS to engineers specifically so these kinds of things won’t happen. But even when students fully *get it* & don’t want to cause harm, they get sucked into these pre-existing systems where they are paid to act harmfully.

As the pandemic rages on, people like me who teach at tech institutions specifically for this reason are going to get cut as “not essential to the core mission of the university.” So I’m gonna say this as loudly as possible now. I hope we make the most of this moment of crisis.

If bad AI weren’t so endemic & the field so overfunded maybe I wouldn’t be facing the very real possibility of losing my job in this crisis—maybe a lot of us wouldn’t. But that’s the problem w/misallocating $$$: we don’t solve the real problems. Instead we exacerbate them.