iiinova — where your digital self keeps evolving

Prolific @joinprolific.bsky.social · 10/10/2026

Thanks to researchers who stopped by our booth at TMRE in Denver this week, shared a controversial research opinion, and picked up merch! We also enjoyed presenting our in-house research on why giving AI models more information makes synthetic data less accurate. #TMRE #MarketResearch #AcademicSky

030

Prolific @joinprolific.bsky.social · 03/10/2026

Our in-house research team found that giving models more information makes synthetic data less accurate. At TMRE, Eric Mercadante (Research Scientist) will discuss why, and what it means for researchers. Be sure to also meet our team and grab swag at Booth #412. #TMRE #MarketResearch #AcademicSky

040

Prolific @joinprolific.bsky.social · 02/10/2026

It was good to meet researchers at ACR Chicago last week! Simon Jones presented on agentic incidence and data quality in online samples, our Beat the Bot game had people spot AI-written answers to win $50 (8 got a perfect score), and we talked research challenges with attendees. #ACR #AcademicSky

031

Prolific @joinprolific.bsky.social · 22/09/2026

We're heading to ACR 2026 in Chicago. Booth #2, Sept 25-26. Talk to us about your research challenges and hear from Simon Jones on agentic incidence, Sept 25 from 8:20 AM. You can enter our raffle too for a chance to win $500 in Prolific credits. #ACR #ConsumerResearch #AcademicSky

020

Prolific @joinprolific.bsky.social · 09/09/2026

In 24 hours, @andrewgordon.bsky.social and Becca Richards will be diving into what makes a research platform trustworthy. We’ve already had great questions around data quality which we'll answer live tomorrow, 12 PM EDT / 5 PM BST. 🔗 riverside.com/webinar/regi... | #AcademicSky #ResearchIntegrity

020

Prolific @joinprolific.bsky.social · 08/09/2026

Has AI been a net positive or negative for survey research? In episode 1 of The Signal, @davmicrot.bsky.social (Microsoft Research, AAPOR) joins @andrewgordon.bsky.social on where AI helps, where synthetic data breaks down, and why transparency beats standards. #AcademicSky #ResearchIntegrity

140

Prolific @joinprolific.bsky.social · 02/09/2026

MTurk's full closure has researchers rethinking online recruitment. Join @andrewgordon.bsky.social and Becca Richards for a fireside chat on participant vetting, platform design, and data quality, backed by years of independent research. 🗓️ Thu Sept 10, 12:00 PM EDT / 5:00 PM BST. #AcademicSky

130

Prolific @joinprolific.bsky.social · 26/08/2026

New research: We tested whether LLM personas can reproduce real survey data, benchmarked against nearly 1,000 real Americans, from zero-context to full digital twins. Richer personas made polls worse. What matters is how you ask - LLMs predict opinion, rather than measure it. #AcademicSky #Research

170

Prolific @joinprolific.bsky.social · 05/08/2026

We're heading to #APA2026 in Washington, D.C., Aug 6-8. 📍 Catch Andrew Gordon's talk "Not All Samples Are Created Equal," comparing data quality across 12 major online sample providers (N = 4,800). Thursday, Aug 6 @ 10-11 AM, Room 147B. Find us at Booth #461 all week. #AcademicSky

030

Prolific @joinprolific.bsky.social · 30/07/2026

Today at Quirk's, @andrewgordon.bsky.social and Asana's Lizzy Munro wrapped up their session on what really determines quality. Backed by a head-to-head comparison study, they shared that participant panel choice remains key, along with tips to maintain research quality. #ResearchIntegrity

040

Prolific @joinprolific.bsky.social · 30/07/2026

AI agent activity in online surveys is far less common than assumed. Across 10 sample providers, agents showed up almost only on MTurk (11-16%), elsewhere at or below 1%. @andrewgordon.bsky.social presents findings today at Quirk's, NYC. Connect if you're there 📄 #AcademicSky #ResearchIntegrity

121

Prolific @joinprolific.bsky.social · 29/07/2026

You can also grab some free merch to take home. That includes chocolate bars that'll tempt you to snap your research time in half!

110

Prolific @joinprolific.bsky.social · 29/07/2026

👋🗽 If you're at Quirk's in NYC today and tomorrow, come find us at Stand Mini A to talk about data quality in online research. #QuirksEvent #AcademicSky

110

Prolific @joinprolific.bsky.social · 28/07/2026

Recent findings show AI agent contamination in online research is far less prevalent than widely assumed. @andrewgordon.bsky.social presents the data this week at Quirk's, with Lizzy Munro (Asana) on what data integrity looks like in practice. 11:15 AM - July 30, Room 1. #AcademicSky #QuirksEvent

020

Prolific @joinprolific.bsky.social · 27/07/2026

@kingscollegelondon.bsky.social Open Research Summer School: £100 to raffle winner Maddy Devonshire (@bristoluni.bsky.social), and a roundtable where our own Senior Product Manager Emilio joined experts to discuss how researchers can keep generating trustworthy data in the age of AI.

100

Prolific @joinprolific.bsky.social · 27/07/2026

EASP (@easp2026.bsky.social): £100 each to five lucky raffle winners, amid poster sessions packed with studies citing Prolific.

100

Prolific @joinprolific.bsky.social · 27/07/2026

BEE UK: £500 to Rebecca Heath (@cam.ac.uk), our Early Career Award winner, after Dr. @andrewgordon.bsky.social took the stage to talk data quality in modern research.

100

Prolific @joinprolific.bsky.social · 27/07/2026

Over the past few weeks, we've had the pleasure of awarding over £1000 in Prolific credit to researchers shaping the future of their fields. A look back at BEE UK, EASP, and King's Open Research Summer School 🧵 #AcademicSky #PhDSky #Research

110

Prolific @joinprolific.bsky.social · 15/07/2026

If you're evaluating research platforms right now, here’s what it looks like when you set up on Prolific. Create an account, connect your preferred tool via link or API, access 300K+ verified participants and 300+ screeners, set your pay, and publish. #AcademicSky #Research

110

Prolific @joinprolific.bsky.social · 08/07/2026

Bring your existing survey link (Qualtrics, SurveyMonkey, Gorilla, Typeform, or whatever you use), set your audience with 300+ free demographic screeners, set your payment (our minimum is $8.00/£6.00 an hour, calculated for you), and publish.

100

Prolific @joinprolific.bsky.social · 08/07/2026

Create an account at prolific.com (no contract or subscription needed), verify your organization or academic affiliation, and add billing details. Pay-as-you-go, no minimum spend.

100

Prolific @joinprolific.bsky.social · 08/07/2026

MTurk closes to new customers on July 30. We want to make moving your research over to Prolific as easy as possible, starting with $50/£50 toward your first study. Here's how to get started on Prolific in 15 minutes ⬇️ #AcademicSky #MTurk

140

Prolific @joinprolific.bsky.social · 08/07/2026

Another spotlight was "Quantifying Frontier LLM Capabilities for Container Sandbox Escape," co-authored by Jerome Wynne with former AISI colleagues - finding that when vulnerabilities are present, frontier models can break out of their sandboxes. 📎 openreview.net/forum?id=19AbP986bv

100

Prolific @joinprolific.bsky.social · 08/07/2026

Big day for our AI research team at #ICML2026 in Seoul 🧪 Prolific's Nora Petrova and John Burden presented "Pressure Reveals Character" - a benchmark of 904 alignment scenarios testing how 24 frontier models behave under realistic pressure. 📎 arxiv.org/abs/2602.20813 | @icmlconf.bsky.social #MLSky

The first print page of the ICML spotlight paper titled "Pressure Reveals Character: Behavioural Alignment Evaluation at Depth."

The second print page of the ICML spotlight paper titled "Pressure Reveals Character: Behavioural Alignment Evaluation at Depth."

230

Prolific @joinprolific.bsky.social · 06/07/2026

📊 Conjointly's 2026 panel experience survey ranked Prolific #1 across participant satisfaction, earnings, and monthly survey completions out of 11 panels. 98% rated our platform easy to use. 96% would recommend us! #MarketResearch #Research #AcademicSky

110

Prolific @joinprolific.bsky.social · 26/06/2026

We're thrilled to share that Prolific is joining as an industry research partner on a new AI4CI Hub grant project 🎉 "Learning Together: Designing Inclusive Human-AI Teams for Collective Decision-Making", led by Dr. Zheng Yuan at @sheffielduni.bsky.social. More details below ↓ #AcademicSky #MLSky

130

Prolific @joinprolific.bsky.social · 08/06/2026

Testing 24 frontier models, they find evidence for a unified "g-factor" of alignment: models tend to be aligned as a whole or not at all. Combined with HUMAINE, the interactive leaderboard is already surfacing some interesting model behavior.

100

Prolific @joinprolific.bsky.social · 08/06/2026

Congratulations to our AI researchers Nora Petrova and John Burden whose paper has been accepted as a spotlight paper at ICML 2026 🧪🎉 "Pressure Reveals Character" tests how frontier models behave under realistic pressure, not just what they say they'd do. @icmlconf.bsky.social #MLSky #Research

A cover photo showing two author names and headshots of the research paper titled "Pressure Reveals Character: Behavioural Alignment Evaluation at Depth" with an indicative label that it has been accepted at the ICML 2026 conference. A coloured deep purple image on the right side containing the Prolific and ICML logos.

150

Prolific @joinprolific.bsky.social · 05/06/2026

Longitudinal studies have always been a pain to run online. Every wave a separate study, allowlists by hand, attrition tracked in spreadsheets. You can now do it all in one organized flow on Prolific. #AcademicSky #BehavioralScience #Research

131

Prolific @joinprolific.bsky.social · 30/05/2026

The Prolific tote bags are a hit. Hope to see a few more of you before the day's end!

030

Prolific @joinprolific.bsky.social · 30/05/2026

If you're at the @psychscience.bsky.social (APS) Annual Convention and haven't stopped by the Prolific booth yet, we'd love to hear from you. Head over to booth #5 in the exhibit hall to find out how we can support your behavioral research. #APS26BCN #AcademicSky #PsychSciSky #Research

160

Prolific @joinprolific.bsky.social · 30/05/2026

Dr @andrewgordon.bsky.social presented "Not All Samples Are Created Equal" at #APS26BCN this week - the most comprehensive comparison of data quality across online sample providers to date. Critical work for researchers making platform decisions. @psychscience.bsky.social #AcademicSky #Research

060

Prolific @joinprolific.bsky.social · 15/05/2026

Fraud in online samples is at an all-time high, and platform choice matters. This afternoon at AAPOR, @andrewgordon.bsky.social presents "Not All Samples Are Created Equal," the largest benchmarking study of data quality across sample providers. 🕓 4:00 PM 📍 Beaudry B, Lobby Floor #AAPOR #Research

020

Prolific @joinprolific.bsky.social · 14/05/2026

Be sure to say hello to the Prolific team at booth #105 if you're at AAPOR this week! We're available throughout the conference to show you how we uphold online data quality, support longitudinal research, and tackle the growing challenge of survey fraud. @aapor.bsky.social #ResearchIntegrity

020

Prolific @joinprolific.bsky.social · 14/05/2026

Can AI replace human respondents? At AAPOR, @andrewgordon.bsky.social presented a benchmark of synthetic survey data against real human polling across multiple question domains, subgroups, and levels of specification. More from Andrew this Friday. @aapor.bsky.social #AcademicSky #Research

030

Prolific @joinprolific.bsky.social · 08/05/2026

Andrew is also joined by Stephanie Eckman and Frauke Kreuter in the opening day workshop "Human-Centered Data Collection for Machine Learning: Lessons from Survey Research." This workshop applies established principles from survey methodology to improve data labeling practices.

100

Prolific @joinprolific.bsky.social · 08/05/2026

→ Not All Samples Are Created Equal: A Comprehensive Comparison of Online Sample Data Quality The first public presentation of our large-scale benchmarking of AI agent prevalence and data quality across the online sampling ecosystem. osf.io/preprints/ps...

110

Prolific @joinprolific.bsky.social · 08/05/2026

→ Impact of Incentive Structure and Sampling Interval on Participant Retention In Online Longitudinal Studies New findings on how pay and session frequency affect participant retention in longitudinal studies, with some surprising results. osf.io/preprints/ps...

100

Prolific @joinprolific.bsky.social · 08/05/2026

→ Can AI Replace Human Respondents? A Systematic Benchmark of Synthetic Survey Data Against Ground Truth Polling Done in collaboration with the Outset team, examining how the depth of synthetic persona specification impacts the accuracy of synthetic data for polling.

100

Prolific @joinprolific.bsky.social · 28/04/2026

The Prolific team at the booth ran demos and our “hot takes” raffle activation throughout the main conference. This was a great chance to show the crowd what human evaluation infrastructure really means in practice.

A photo of a large crowd surrounding the Prolific booth at the ICLR conference.

A medium close-up photo of a Prolific team member interacting with a visitor at the Prolific booth at the ICLR conference.

100

Prolific @joinprolific.bsky.social · 28/04/2026

We hosted our exclusive Prolific x AI Circle Lights Over Rio dinner on Sugarloaf Mountain, bringing together researchers from across the AI community over Brazilian cuisine and breathtaking views.

A photo of guests socializing outdoors on a sunny day.

A photo of Rio de Janeiro city lights at night.

100

Prolific @joinprolific.bsky.social · 28/04/2026

Nora Petrova and John Burden from our sciences team presented “The Missing Red Line” at the ICBINB ("I Can't Believe It's Not Better") workshop. Their research finds that commercial system prompts can override model safety training: openreview.net/forum?id=ev5...

A poster showing findings from the research paper titled "The Missing Red Line" by Prolific's sciences team.

100

Prolific @joinprolific.bsky.social · 28/04/2026

Enzo and John gave an expo talk on building evaluation infrastructure for frontier AI to a packed room.

A photo of a speaker stood on stage next to a blue presentation screen that reads "Your models are outgrowing your evaluations."

A photo of a large audience at an expo talk.

100

Prolific @joinprolific.bsky.social · 28/04/2026

Our Sr. AI Research Engineer John Burden presented the “Formalising Human-in-the-Loop” poster, co-authored during his time at CFI at @cam.ac.uk (work led by Maurice Chiodo as part of the Ethics in Mathematics Project): openreview.net/forum?id=KR8...

A photo of the first page of the research paper titled "Formalising Human-in-the-Loop: Computational Reductions, Failure Modes, and Legal-Moral Responsibility." The photo includes the author names, paper abstract, and introduction.

100

Prolific @joinprolific.bsky.social · 28/04/2026

On the first day of the conference, we co-hosted an Afternoon Tea session with Encord and @modal-labs.bsky.social, gathering researchers and engineers for informal conversations over tea and coffee. ☕️

100

Prolific @joinprolific.bsky.social · 28/04/2026

Nora Petrova and Enzo Blindow of Prolific's sciences team presented HUMAINE, our peer-reviewed framework for demographically stratified LLM evaluation. Paper and live leaderboard: www.prolific.com/humaine?utm_...

A poster showing findings from "Unpacking Human Preference for LLMs", a research paper by Prolific's sciences team.

110

Prolific @joinprolific.bsky.social · 28/04/2026

Five days in Rio saw a lot of great research, scientific conversations, and connections at ICLR 🧪 Here's an #ICLR2026 highlights thread. @iclr-conf.bsky.social | #MLSky #AcademicSky #ArtificialIntelligence #Research

A group photo of dinner guests captured at night in Rio de Janeiro, during the ICLR Conference.

120

Prolific @joinprolific.bsky.social · 26/04/2026

Tomorrow our sciences team are presenting "The Missing Red Line" at the #ICLR2026 ICBINB workshop which focuses on where LLMs need to improve and what rigorous failure audits reveal. Our researchers tested 8 frontier models in scenarios where commercial objectives conflict with user safety. 1/4 🧪

ICLR conference poster image for the paper titled "The Missing Red Line" that shows a summary and key findings from the research by Prolific's sciences team. The thread talks about these findings as the paper will be featured at an upcoming workshop.

120

Prolific @joinprolific.bsky.social · 25/04/2026

On our way to Sugarloaf Mountain for dinner with the AI research community ⛰️ #ICLR2026

030

Prolific @joinprolific.bsky.social · 25/04/2026

Nice to see a packed room for our #ICLR2026 expo talk! Enzo, VP of Data & AI, and John, Sr. AI Research Engineer, shared lessons learned from building evaluation infrastructure for frontier AI and research. They talked about what works and surfaced the failures that matter. #ICLR #MLSky #Research

120