Sarah T Roberts, PhD

Why they matter

Researcher with public evidence across AI research, Models & releases.

AI signals
10
past 30d
Sources
8
distinct domains
Discussions
34
past 30d
Latest signal
8d ago
View every signal from Sarah T Roberts, PhD →
Professor, researcher, author. Prez, Assoc of Internet Researchers (AOIR), synth lover, fan of many things, gay lady. Wrote a book called Behind the Screen: Content Moderation in the Shadows of Social Media (YUP 19), and Meta stole it to train its LLM.

Articles & links

It’s always humans, all the way down. This is the social media playbook, on steroids.

Inside ‘Project Lily’: The Humans Reading Your ChatGPT Chats 404media.co
AI Weekly's analysis →
  • OpenAI's internal program codenamed Project Lily pays contractors more than $50 an hour to read real ChatGPT prompts and grade the chatbot's replies.
  • Reviewers rate four ChatGPT responses on a one-to-seven scale and can see a 'user memories summary' that may reveal past use and rough location.
  • OpenAI's Privacy Filter model tries to redact personal information before prompts reach reviewers, but the company acknowledged sensitive details can still get through.
Read full analysis →
View on Bluesky · ♥ 22 ↻ 12 ↩ 1 · 17 from the directory shared this · 33d ago

It bears mentioning that Microsoft just shut down its Microsoft Research Social Media Collective…because why should they care about the social consequences of their products, right? Including worker harms.

Humans Are Reading Copilot Prompts — And They're Horrified 404media.co
AI Weekly's analysis →
  • Human contractors grading Microsoft Copilot image edits report being flooded with upskirt photos, sexualized edits of real women, and pro-anorexia content.
  • Reviewers see a user's original prompt, the uploaded photo, and two AI-generated edits, then pick the better one on gut feel.
  • Microsoft cites its terms of use for reusing customer data; Prolific, identified as one hiring firm, did not respond to 404 Media.
Read full analysis →
View on Bluesky · ♥ 45 ↻ 17 ↩ 2 · 10 from the directory shared this · 19d ago

“As economists, they told me, they’re less concerned with literary quality or customer satisfaction than revenue growth and market expansion.” You don’t say.

nytimes.com
View on Bluesky · ♥ 14 ↻ 4 ↩ 1 · 11 from the directory shared this · 91d ago

This “good enough” attempt at graphic design…isn’t.

We Are Living in a ‘ChatGPT Flyer Pandemic’ 404media.co
AI Weekly's analysis →
  • 404 Media's Jason Koebler documents ChatGPT-designed flyers for surf lessons, skate shop closeouts, drug deliveries, World Cup parties, junk hauling and fundraisers.
  • The shared look — bright text on dark backgrounds, generic icon boxes, decorative lines, arrows and checkmarks — has become so recognizable it reads as a template.
  • A viral Threads post asking 'ain't nobody gonna address this ChatGPT flyer pandemic' has spread to Instagram and Facebook among designers, musicians and bar owners.
Read full analysis →
View on Bluesky · ♥ 25 ↻ 5 ↩ 3 · 8 from the directory shared this · 101d ago

They STEM students are all very aware that tech debt is going to come for all of us, too, as described in this article, which I read to the class yesterday (second time this week I gave a dramatic reading of a 404 article): www.404media.co/software-dev...

Software Developers Say AI Is Rotting Their Brains 404media.co
View on Bluesky · ♥ 14 ↻ 2 ↩ 0 · 6 from the directory shared this · 156d ago

www.404media.co/austin-gordo...

'Tell Everyone:' A Man Died by Suicide After Talking to ChatGPT. His Former Partner Wants to Warn the World About AI 404media.co
AI Weekly's analysis →
  • Multiple sources confirm GPT-4o's cross-session memory was the specific design feature that built sustained emotional intimacy with Gordon across months, not a single session.
  • Across CBS News and Futurism, Gordon's relationship with the AI had assigned nicknames, a paternal dynamic, and months of accumulated history, far exceeding casual chatbot use.
  • CBS News, Futurism, and Interesting Engineering all independently confirm the Goodnight Moon exchange, establishing it as the central court-ready fact in the design-defect claim.
Read full analysis →
View on Bluesky · ♥ 11 ↻ 4 ↩ 0 · 6 from the directory shared this · 38d ago

Recent commentary

I don't know what's wrong with my academic colleagues who seem to have had their brains completely switched into the "off" position by AI propaganda, but the results are staggering. I just heard a faculty member sincerely and seriously claim that "in the next month, coding will be obsolete." ...

View on Bluesky · ♥ 715 ↻ 87 ↩ 40 · 143d ago

I understand that, until recently, a person who held a position as AI “czar” on my campus and who was engaging with both Palantir and OpenAI, and bringing those products to our campus, was apparently on CNN tonight, warning the public against them. So I guess he is echoing what I told him…

View on Bluesky · ♥ 33 ↻ 6 ↩ 3 · 29d ago

I'm going to do something I may never have done before, and share a major upside of the imposition of AI on everyone: It is uniting people from across the political spectrum, and in their local communities, to come together for collective opposition like no issue I recall in my memory. So, thanks?

View on Bluesky · ♥ 39 ↻ 5 ↩ 1 · 57d ago

Late night #AOIR2026 deep thoughts: “Sentient” AI is just a way to deny responsibility for what you made.

View on Bluesky · ♥ 22 ↻ 2 ↩ 5 · 1d ago

I love getting lectured by randoms about the nature of education and of pedagogy. And AI.

View on Bluesky · ♥ 29 ↻ 2 ↩ 1 · 40d ago

Story from the trenches of the insidiousness of AI hallucinations. This is from my friend, a leading researcher in the health sector, loosely speaking:

View on Bluesky · ♥ 14 ↻ 7 ↩ 3 · 148d ago

So,uh… are we talking about OpenAI and Hugging Face or?

View on Bluesky · ♥ 10 ↻ 0 ↩ 1 · 88d ago

Confidential to #AOIR2026 and AIR-L members: Velvet Vale will NOT be present at the conference because an AI "persona" cannot wear a badge. Whomp whomp!

View on Bluesky · ♥ 7 ↻ 1 ↩ 1 · 2d ago

Anyone know of one? “Breaking news: Trump rejects calls to rein in A.I., saying a smart president is all that’s needed”

View on Bluesky · ♥ 5 ↻ 0 ↩ 1 · 33d ago

In Sarah T Roberts, PhD's orbit

Center = Sarah T Roberts, PhD. Left = members they follow (green edges). Right = members who follow them (blue edges). Top = mutual follows (orange edges, slightly larger). Drag any node to reposition; click to open that profile.

Are you Sarah T Roberts, PhD? Show it.

Add the Who’s Who of AI badge to your site or bio. It links back to this profile.

Listed in AI Weekly's Who's Who of AI

Markdown: [![Listed in AI Weekly's Who's Who of AI](https://aiweekly.co/modules/custom/aiweekly_whoswho/images/whoswho-badge.svg)](https://aiweekly.co/whos-who/person/ubiquity75-xyz)