Notices by Carl T. Bergstrom (ct_bergstrom@fediscience.org)

Tonight I had one of the most glorious birding moments of my life. A decade ago I first saw pinyon jays in the Colorado mountains. Ever since I've looked for them in the mountains around Santa Fe on my frequent visits there. I can usually find them, but it takes time and I see just a few.

After massive commercial deforestation and climate-change driven destruction of the pinyon forests they rely on, their numbers are plummeting and they are being considered for the endangered species list. 1/2

Tonight after a great day working with colleagues at the Santa Fe Institute, Joe Bak-Coleman went for a hike. As we headed back to the car, A huge flock of birds headed up Santa Fe Canyon; I knew their squeaky caws before I even saw them. It was a group of at least 100 pinyon jays heading to roost. I've never seen anything like it. I am so happy they are thriving here. 2/2

#birdwatching #birdphotography #corvids

Another Elsevier paper with obvious AI-written text.

“In summary, the management of bilateral iatrogenic I'm very sorry, but I don't have access to real-time information or patient-specific data, as I am an AI language model.”

https://www.sciencedirect.com/science/article/pii/S1930043324001298

Frames within frames, Barcelona.

#photography

Douglas Allchin, Jonathan Osborne, and I have a new paper out about teaching science in this age of online misinformation.

The idea is simple: we tell people to trust the science, but we rarely teach students about the social processes that make it trustworthy.

https://carlbergstrom.com/publications/pdfs/2024JCST.pdf

My daughters and I are feeding cows 10,000 pounds of macadamia nuts each and that's what makes us so relatable and makes it so clear we don't need to be taxed at 95+ percent of our income.

“countless professors, when asked to perform a peer review for a journal, fob the work off to their hapless grad students”

Having been through the fact-checking process at the Times repeatedly, I can’t understand how nonsense like this gets through.

I can only speculate the perhaps these days their fact-checkers fob the work off to hapless interns.

https://www.nytimes.com/2024/01/04/opinion/plagiarism-academia-claudine-gay.html?smid=nytcore-ios-share&referringSource=articleShare

It’s gonna be just great when the people who keep FOIAing massive volumes of email correspondence from public university faculty in order to dragnet for wokeness start running those collections through a large language model that fabricates messages.

https://futurism.com/the-byte/google-gmail-tool-hallucinating-emails

Unclear signage led to crocodile attacks, study finds.

(photo: Parks Australia)

Ah yes, of course. I've been trying to figure out who the fuck to blame.

https://thehill.com/policy/energy-environment/4057045-catch-22-scientific-communication-failures-linked-to-faster-rising-seas/

Here is a must-read post from children's author Maggie Tokuda-Hall on how Scholastic offered to publish her book — all she had to do was remove all mention of racism.

Sure, they're banning books in Tennessee and Texas. But it's not just the books that get published and then banned from the library. It's all the books that don't get published in the first place.

Those banning books know publishers like Scholastic pull this cowardly bullshit. It's their game plan.

https://www.prettyokmaggie.com/blog/2023/4/11/scholastic-and-a-faustian-bargain

Here is your must-read article for the day, a profile of @emilymbender, and her efforts to deflate the ridiculous hype around large language models such as ChatGPT.

It's also about the people who are behind that hype, and about what their way of thinking has the potential to do to us.

It's worth reading all the way to the end.

https://nymag.com/intelligencer/article/ai-artificial-intelligence-chatbots-emily-m-bender.html

Meta. OpenAI. Google.

Your AI chatbot is not *hallucinating*.

It's bullshitting.

It's bullshitting, because that's what you designed it to do. You designed it to generate seemingly authoritative text "with a blatant disregard for truth and logical coherence," i.e., to bullshit.

Please read this and everything else Cory Doctorow
@pluralistic writes.

https://pluralistic.net/2023/02/02/nonbinary-families/#red-envelopes

This should be obvious, but having an algorithm that behaves that way is a DELIBERATE CHOICE.

It would be easy enough, for example, to implement basic sentiment analysis so that the algorithm doesn't boost content that you have reacted negatively to in the future.

Musk is playing it both ways. He keeps the algorithm that boosts inflammatory content and drives the online conflicts that draw views and clicks, while pushing the blame for this off onto the individuals involved.

That sucks.

I've been speaking and writing lately about how tech companies try to shift the discussion on misinformation, polarization, harassment, etc., away from the systems and structures that are inherently toxic and toward questions of individual behavior.

This way they can blame their own users for any pathology and steer clear of calls for systemic change.

Today Elon Musk has come through with a perfect illustration for my future talks.

#socialmedia #twitter #ElonMusk

I don't even know where to start with this one. Apparently the mental health service kokocares.org ran a multi-day trial on 4000 users without their informed consent, in which koko's "peer supporters" used ChatGPT to craft their responses to the people they were providing with mental health support.

https://www.businessinsider.com/company-using-chatgpt-mental-health-support-ethical-issues-2023-1

Co-founder Rob Morris tweeted a description of the trial in this thread, in which he explicitly describes the activity as an experiment.

https://twitter.com/RobertRMorris/status/1611450197707464706?s=20&t=ndcUoODJItgoTH8mlf8a9Q

Morris later tried to walk back his original claims and suggested that he had informed consent from all participants.

But it appears from his own claims that only the "peer supporters" were informed.

According to Business Insider, `Koko users were not initially informed the responses were developed by a bot, and "once people learned the messages were co-created by a machine, it didn't work," Morris wrote on Friday.'

What makes this different than your average tech-company-does-very-shitty-thing-using-AI-without-informed-consent story is that the co-founder was so oblivious to any notion of research ethics, consent, mental health ethics, etc., that he wrote a long thread describing this experiment and was caught completely off guard when people reacted with shock.

I think this example tells us a lot about whether we can trust tech to act ethically: looks like they don't know how, even when they want to.