Post

i teach college writing and information literacy and i make a point of doing a unit on bad faith arguments. i drill it into the students that they do not have to acknowledge all sides of an argument in order to discuss it well or sound informed.

I get that, but only telling one side is how we got Fox News.

I wonder if people that say that think AI is unlikely to go away & want to know guardrails that you as an expert hope for? I do use that "easier to critique than build" line only when someone posts advice to Dems on-line when they appear to have never organized a potluck or a PTA meeting.

Once, when pitching an article about the grim reality of content moderation and data labelling in the AI industry, the editor came back to say that I "shouldn't put my finger on the scale" and that there should be space given to the potential benefits (of traumatizing, poorly paid labour). 🫠

"Any potential benefits are beyond the scope of this paper."

argh

It’s a piece that isn’t balanced but is falsely written as one that is. Take out the concern piece that supposedly balances the piece and there is nothing there.

I've noticed a real strand of hit dogs hollering when it comes to trying to have pro/con discussions about AI in academia. I think there are a lot of faculty who are on the DL about how much they are offloading to AI and it comes out as unwarranted vitriol when someone reminds them of the down sides

This is the closest I will ever come to acknowledging both sides: bsky.app/profile/cv-d...

Yes, all that throat-clearing is so tiresome, especially since most of it isn't backed up by any real data.

There are also academics who frame the task of AI ethics as "making sure that the benefits of AI can be realized while minimizing harms"... no? The claim that any particular tech should exist needs to be the end point of an ethical inquiry, not the starting point! It's not a forgone conclusion...

Option One. Do not repair. Cost: $0. Outcome: Boat sinks at dock. Option Two: Minimal repairs. Cost: $65k. Outcome: use of boat for another 6months. Option Three: Full repair. Cost: $200k. Outcome: use of boat for 12yrs (w/ ongoing routine-only repair & maintenance). 2/2

uggggh

It's also just such a lazy point for peer reviewers to make. Like, that's the thing you write when you put off the peer review until the last minute and skimmed the abstract but wanted to pretend you have something to say. It's so...gross. Normalized, but gross.

I have a couple of these. One on disability studies (we’re not supposed to critique anything in the framing of the authors who made the call), and one management of language difference – in which I didn’t understand how badly they needed me to toe the line. (and, to be fair I fucked up the resubmit)

Imagine a historian starting an article with, "While we must accept that IG Farben was the producer of many important chemicals that improved agriculture and hygiene..."

It's imperative that we assess both sides. It's not imperative that we conclude that one side isn't so wrong that it's not worth discussing.

bsky.app/profile/oliv...

“Your critique is valid for the current state of thing but not about this possibility a company sold me on so your argument is invalid” is just not something an editor should allow through peer review

It’s all harms though… is the thing. Even if it delivered on its promises it would only make things worse, both for individuals and corporations. It’s such an obvious scam it hurts :(

@tarcio.bsky.social

Will add that a lot of comments here reflect the sort of mandates journalists have for every story where there are reasons to bring in multiple voices and perspectives. I don’t see academic or other research needing that. It can even weaken it.

Balance for sake of balance has no place in intellectual arguments. Yes.

Some of the pros really help disabled access to technology for me, personally. However, I don't think they are worth the cons.

While it is true that we are likely to live in a world in which powerful people, in particular the employers of the graduates we are producing now, will remain convinced of the benefits of neural network based AI for at least the next five years, ...

i think lit review is tangentially to blame for this both-side-ism! i think it should be possible to make a case for how AI sucks without having to explain why it's cool

Recognizing which situations affect each on a systematic level. The problem with all of the ones you have mentioned. You also divulge where they actually would occur. That is not just keeping you accountable but keeping ai accountable

It isnt a both sides like op is claiming but actually a way to show you have done the analysis of different arguements and can critique them with your evidence

🤣

😂🙌🏾

Because, you know, there aren't numerous pieces just focusing on the benefits of AI. Well, more than just focusing on benefits - literally frothing at the mouth to extol the industry and "progress."

When the entire premise is based on the unethical theft of the creative commons, (and it worsens energy demands), then it needs no further debate.

It’s not like focusing on the negatives will hurt the AI’s feelings.

They need to be taught the environmental harms of AI. (Oh, we teach them media literacy.) .... .... I am VERY close to asking for her to be kicked out. Like, ma'am, you tried. Every argument you had was either specious, or arguing for the very things we're against.

I’m tired of no one starting with the fact that there is road map to make this shit financial viable. we’re headed towards another financial crisis

Primarily by wasting no time on them. But that is not the same as what is being proposed here. Example…

Thinking about both-sidesing other issues: "cigarettes help you relax and lose weight, but did you know they also cause cancer? It's so complicated because of the benefits and harms"

nobody has ever proven a single supposed benefit so I wouldn't bother stating them

Hell yes. It’s wild to see

Call it the dril principle - elegantly stated in both their Isis and drunk driving posts.

Most of the pros are hypothetical: people speculate about how useful it *could* be, while minimizing how harmful it actually already *is*. And bluntly, academic honesty generally does not require going over hypothetical alternate realities.

I was just saying yesterday that if my role is to be the only AI pessimist in the room reminding folks of the actual harm they are willing accept because "AI is useful," then so be it.

Some of this is just too on the nose.

To harness the health benefits of dietary radium ... The asbestos revolution is transforming home safety ... Lead pipes democratize water ... Thalidomide is shaping the future infants ...

To be fair, tobacco still does that sadly. Some research linked to from here, if you're interested. olivia.science/before

2 years of seeing my friends and colleagues in the climate movement doing this has radicalised me

Is there a societal upside to the negative economic impacts brought on by the bubble burst when it becomes clear to AI financiers that there’s no increased income stream expected from AI? I doubt it’ll provide a service I’ll ever pay for. I won’t even pay for cloud service to Apple or Microsoft.