emilymbender

emilymbender, 1 year ago

“The group working on ethics that Google pledged to fortify is now disempowered and demoralized, the current and former workers said.”

>>

emilymbender, 1 year ago

Google’s leaders decided that as long as it called new products “experiments,” the public might forgive their shortcomings, the employees said.

➡️We don’t tolerate “experiments” that pollute the natural ecosystem and we shouldn’t tolerate those that pollute the information ecosystem either.

>>

emilymbender, 1 year ago

“Silicon Valley as a whole is still wrestling with how to reconcile competitive pressures with safety.”

➡️Are they though? It seems to me that those in charge (i.e. VCs and C-suite execs) are really only interested in competition (for $$).

“But ChatGPT’s remarkable debut meant that by early this year, there was no turning back.”

➡️False. We turned back from lead in gasoline. We turned back from ozone-destroying CFCs. We can turn back from text synthesis machines.

>>

emilymbender, 1 year ago

“On the same day, [Google] announced that it would be weaving generative AI into its health-care offerings.”

➡️ 🚨🚨🚨

“Employees say they’re concerned that the speed of development is not allowing enough time to study potential harms.”

➡️Employees are correct.

>>

emilymbender, 1 year ago

“One former employee said they asked to work on fairness in machine learning and they were routinely discouraged — to the point that it affected their performance review.”

➡️Not a good look, Google.

“But now that the priority is releasing generative AI products above all, ethics employees said it’s become futile to speak up.”

➡️And it shows…

>>

emilymbender, 1 year ago

“When Google’s management does grapple with ethics concerns publicly, they tend to speak about hypothetical future scenarios about an all-powerful technology that cannot be controlled by human beings”

➡️So tempting to focus on fictional future harms rather than current real ones.

Thank you, @daveyalba for this reporting.

emilymbender, 1 year ago

@timnitGebru @xriskology

Also great for understanding what the #TESCREAL bundle of ideologies is, how they connect, and why any serious work towards improving things for people on this planet should be very clearly distanced from any of that.

>>

emilymbender, 1 year ago

@timnitGebru @xriskology

And just to be very clear: If your work has been exposed as pointing to eugenicist or otherwise racist underpinnings, it's not enough to just "disavow".

If you want to break that connection, you've got to do the work: read those who have been documenting the harms, understand how those harms relate to the work you were pointing to and interrogate how the concepts you've been drawing on could mean your work is perpetuating harm.

>>

emilymbender, 1 year ago

#AIhype isn't the only problem, for sure, but it is definitely a problem and one that exacerbates others. If LLMs are maybe showing the "first sparks of AGI" (they are NOT) then it's easier to sell them as reasonable information access systems (they are NOT).

>>

emilymbender, 1 year ago

If (even) the people arguing for a moratorium on AI development do so bc they ostensibly fear the "AIs" becoming too powerful, they are lending credibility to every politician who wants to gut social services by having them allocated by "AIs" that are surely "smart" and "fair".

>>

emilymbender, 1 year ago

If the call for "AI safety" is couched in terms of protecting humanity from rogue AIs, it very conveniently displaces accountability away from the corporations scaling harm in the name of profits.

>>

emilymbender, 1 year ago

It's frankly infuriating to read a signatory to the "AI pause" letter complaining that the statement we released from the listed authors of the Stochastic Parrots paper somehow squandered the "opportunity" created by the "AI pause" letter in the first place.

>>

emilymbender, 1 year ago

Yes, we need regulation. But as we said:

"It is indeed time to act: but the focus of our concern should not be imaginary "powerful digital minds." Instead, we should focus on the very real and very present exploitative practices of the companies claiming to build them, who are rapidly centralizing power and increasing social inequities."

https://www.dair-institute.org/blog/letter-statement-March2023

>>

emilymbender, 1 year ago

@RuthStarkman @UlrikeHahn my comment was for Ulrike. If you're citing something that was published in a closed venue and posted to arXiv, then I think you cite the published/reviewed version and also give the arXiv link.

emilymbender, 1 year ago

@zoeschiffer @caseynewton

Self-regulation was never going to be sufficient, but I believe that internal teams working in concert with external regulation could have been a really beneficial combination.

https://twitter.com/emilymbender/status/1336517038525665280?s=20

>>

emilymbender, 1 year ago

@zoeschiffer @caseynewton

So, what can we do as researchers and otherwise, in this moment?

I have to believe there is value in calling out #AIhype when we see it, in declining to participate in the #AIhype, and in advocating for regulation.

And those who are making buck off of this will try to tell us: You can't regulate this --- that will stifle progress!

To which I ask: Progress towards what, exactly?

>>

emilymbender, 1 year ago

@zoeschiffer @caseynewton

And they will tell us: You can't possibly regulate effectively anyway, because the tech is moving too fast.

But (channeling Ryan Calo here): The point of regulation isn't to micromanage specific technologies but rather to establish and protect rights. And those are enduring.

>>

emilymbender, 1 year ago

@zoeschiffer @caseynewton

I call on everyone who is close to this tech: we have a job to do here. The techcos where the $, data and power have accumulated are abandoning even the pretext of "responsible" development, in a race to the bottom.

At the very least, we should be working to educate those around us not to fall for the hype---to never accept "AI" medical advice, legal advice, psychotherapy, etc.