Hacker News

Top stories

Live mirror
30 storiesupdated just nowView source snapshot
  1. OpenAI GPT–6 Astra breaks Enigma message that has resisted solution since 2005(cryptocellar.org)
    158comments
  2. OpenAI is about to eat Jev's lunch – Arcturus Labs(arcturus-labs.com)
    9comments
  3. Show HN: Drop – a rootless Linux sandbox with gVisor support(droprun.sh)
    5comments
  4. Apple has added persistent 'ads' to iOS, and it's driving users crazy(techradar.com)
    31comments
  5. Jev – a curation of Jev demos on X, tools, skills, and integrations(github.com/amal-david)
    discuss
  6. Can gzip be a language model?(nathan.rs)
    111comments
  7. MiMo v2.6(xiaomi.com)
    452comments
  8. Spymarks, Not Watermarks(brand.io)
    139comments
  9. What Capital Never Told You About Rent(humansontheloop.com)
    1comments
  10. We broke an Over-The-Air update on the ESP32 on purpose(groundrun.io)
    1comments
  11. Line Scan Photos from MUNI Heritage Weekend in San Francisco(lawrence.lu)
    12comments
  12. Quantum information spreading via higher-order operator correlators(arxiv.org)
    discuss
  13. Stanford R&DE Uses AI to Race Swap Students for Advertising(stanfordreview.org)
    36comments
  14. Attention is all you have(alicegg.tech)
    288comments
  15. Transformers Explained Visually(poloclub.github.io)
    79comments
  16. I said no and Apple said yes(dbushell.com)
    444comments
  17. What Sun got wrong(dtrace.org)
    367comments
  18. AMD's random number generator can't generate a 0?(flatassembler.net)
    127comments
  19. A font that reads what you wrote(rohanadwankar.github.io)
    32comments
  20. AI Has No Wisdom and Neither Will You(alexn.org)
    386comments
  21. 9 Ads per Minute: FIFA Cup 26 – "the price of the beautiful game"(bristol.ac.uk)
    172comments
  22. I don't want to read what you didn't write(colinbreck.com)
    366comments
  23. MiMo-v2.6-Pro: Intelligence, Performance and Price Analysis(artificialanalysis.ai)
    47comments
  24. Engineering Memory: On learning to memorize first 100 digits of pi (2024)(gregorygundersen.com)
    29comments
  25. AI coding has made CI a bottleneck, so we reworked ours to keep up(linear.app)
    349comments
  26. What It's Like to Work in One of America's Data Centers(wsj.com)
    36comments
  27. People Training OpenAI's AI Fired for Using AI to Train the AI(404media.co)
    20comments
  28. Verda (Finland) raises $189M in Series B(verda.com)
    27comments
  29. Divide by depth for instant 3D(gabrieloc.com)
    35comments
  30. NASA’s Mars Sample Return mission is dead(science.org)
    348comments

Stanford R&DE Uses AI to Race Swap Students for Advertising

50 pointsby 1h agostanfordreview.org
31 comments
44m agoHN ↗

Really seems like they used it to anonymize the students, but that is a lot less polarizing. Does it matter if they swapped the race? They also swapped all identifying facial features of anyone in the picture. That might have even been the prompt they used in the first place.

I am not a big fan of gen ai, but freak outs about nothingness detract from the real issues. I would be more concerned if they swapped the 3 students into a background or setting they were never in, than use the public setting and swap out the students for ai models.

43m agoHN ↗

Only a bot who can't parse images would reply with this. Look at the pictures, your thesis is totally wrong.

36m agoHN ↗

dang's law: Never attribute to a bot that which is adequately explained by someone not reading TFA

15m agoHN ↗

funny but presumably this law predates hn. is it actually attributable to him?

full disclosure, this user did not read the article.

41m agoHN ↗

That explanation doesn't make sense. If the point is anonymity, why not just AI generate the students entirely? They AI-generated the girl on the right in her entirety, so capability clearly isn't the problem. And the guy in the middle is still very much recognizable.

20m agoHN ↗

I don't understand what you mean? For me, it appears all 3 students have been changed. The 3rd student is changed into a black female. I am not doubting they want to inject diversity (although i didnt say it above) my contention is they likely view the original as a blank background, which they inserted 3 "new" anonymized students into. Its not that they turned 1 person "black", its that they changed ALL 3 PEOPLE entirely

16m agoHN ↗

Given how little the two students on the left changed, I highly doubt the AI made a creative decision to completely randomize only the student on the right. I think it's more likely that they did it in two steps: tweak the faces, and replace the student on the right.

44m agoHN ↗

Is this a race to find the most offensive use of AI?

36m agoHN ↗

Amazing how it's offensive to everyone involved. Both to be replaced and to be represented by an AI rendition.

38m agoHN ↗

It's more offensive to be prettified by an AI than it is to be removed and replaced entirely.

37m agoHN ↗

They went the opposite way from what I assumed from the headline. Also they clearly thought that girl wasn’t slim enough.

3m agoHN ↗

Yeah honestly anyone familiar with marketing and communications departments (and especially those in higher-ed) is used to this sort of chicanery.

This particular instance is somewhat notable in just how egregious it is, but they pick-and-choose students based solely on ethnicity for photoshoots all the time.

14m agoHN ↗

Still just as bad IMO. Whitewashing and blackwashing are both racist.

7m agoHN ↗

It looks more like the image on the right is of a banner and isn't flat.

32m agoHN ↗

Tokenism is awful. Unfortunately, speaking out about it is also frowned upon nowadays. To be perfectly clear, I'm only talking about marketing efforts that intentionally ignore or replace unmarked groups with marked ones. Here, they took a real photo and fucking replaced students that didn't look diverse enough for their tastes. I'd argue that should be considered just as bad as the reverse.

To downvoters: I AM NOT saying there should be less diversity. I am saying that fake and dishonest diversity is just as harmful as any other kind of discrimination. Instead of helping the problem it just creates an alternate version.

31m agoHN ↗

My account is banned, but for whoever has [showdead]:

This is totally self-inflicted.

30m agoHN ↗

Self-inflicted by whom? I really don't think the people in the photo are the problem. I think whoever thought they needed to fucking edit it to create a fake impression of diversity needs to stop. If you want to show that your school has diverse students... then take diverse photos. Don't take a non-diverse photo and then diversify it. At best (there are diverse students, but they weren't in the photo) it's dishonest, at worst (there aren't diverse students) it's literally false advertising.

30m agoHN ↗

So Stanford R&DE subscribes to the Jason Arday school of thought.

28m agoHN ↗

Marketing is about attracting customers.

So the incentives are there to present the image you want to project to attract customers.

21m agoHN ↗

Hey buddy, not everything is about cold calculation. Some of us value culture and integrity. These are going to matter even more in the age of AI not less.

19m agoHN ↗

Advertising is more about selling aspirations rather than reflect objective reality.

To wit: even the original photo was probably staged…

18m agoHN ↗

Anything you do as a university / person reflects your values. This shows a lack of integrity.

14m agoHN ↗

And which exactly customers will be enticed by the modified image?

11m agoHN ↗

Yes, which is why people study marketing to understand revealed preferences in society. Here, they slimmed down the woman on the left. It's well known that skinnier is more marketable. By the same token, we must assume that they think replacing the hispanic man with a black woman is also an upgrade. By contrast, in China, they sometimes do the opposite: https://variety.com/2015/film/news/star-wars-china-poster-co....

20m agoHN ↗

To quote one of the most prolifically insufferable linkposters of HN: "Demographics are destiny."

So, you fill your university marketing and publication department with people who think this is ok and this is the sort of PR faux-pas whoopsie you get.

I'm sympathetic to the argument that this sort of marketing material needs as many races represented as possible so as many people as possible can see it and see someone that looks like them and identify with it but something tells me the university isn't gonna come out and own it like that and even then using AI to make a replacement when you have a student body to pull from is kind of icky.

6m agoHN ↗

I fail to see how these are related whatsoever.

If anything, it adds more fuel to the maga "woke academia" fire.

8m agoHN ↗

I'm not sure which is worse: human marketer instructing AI specifically to make everyone more diverse and attractive or human marketer instructing AI to do basic marketing and AI has internalized to make everyone more diverse and attractive.