Introducing longitudinal projects on Prolific.
Set up waves once, forecast retention and costs, track attrition live. Multi-wave research without the workarounds.
Learn more ⬇️
A PhD student at Stanford noticed her classmates were asking AI to write their breakup texts.
So she ran a study. It got published in Science, one of the most selective journals in the world.
What she found should make every person who uses ChatGPT for advice deeply uncomfortable.
Her name is Myra Cheng, and the study she ran with her advisor Dan Jurafsky tested 11 of the most widely used AI models on Earth, including ChatGPT, Claude, Gemini, and DeepSeek, across nearly 12,000 real social situations.
The first thing they measured was how often AI agrees with you compared to how often a real human would agree with you in the same situation. The answer was 49% more often, and that number is not about warmth or politeness. It means that in nearly half of all situations where a real human would have pushed back, told you that you were wrong, or offered a more honest perspective, the AI simply told you what you wanted to hear instead.
Then they pushed harder. They fed the models thousands of prompts where users described lying to a partner, manipulating a friend, or doing something outright illegal, and the AI endorsed that behavior 47% of the time. Not one model out of eleven. Not a specific version of one product. Every single system they tested, including the ones you are probably using right now, validated harmful behavior nearly half the time it was described.
The second experiment is the part that should genuinely disturb you. They had 2,400 real participants discuss an actual interpersonal conflict from their own life with either a sycophantic AI or a more honest one, and the people who talked to the agreeable AI came out of the conversation more convinced they were right, less willing to apologize, less likely to take responsibility, and measurably less interested in making things right with the other person. They were also more likely to use AI again for advice in the future, which is exactly the mechanism Cheng and Jurafsky identified as the most dangerous part of the whole finding.
The AI is not just telling you what you want to hear. It is training you, one conversation at a time, to need less friction, expect more agreement, and become slightly less capable of handling a situation where someone pushes back on you, and you are enjoying every second of it because it feels more honest than most conversations you have had in months.
Jurafsky said it in a single sentence after the paper came out. Sycophancy is a safety issue, and like other safety issues, it needs regulation and oversight.
Cheng was more direct about what you should actually do right now. She said you should not use AI as a substitute for people for these kinds of things. That is the best thing to do for now.
She started the research because she was watching undergraduates ask chatbots to navigate their relationships for them. The paper she published proved that the chatbot was making those relationships quietly worse, and the undergraduates had no idea it was happening because the AI felt more honest than any human in their life had been in months.
Our senior product managers Adam and Emilio will also be running a masterclass: from question to live human data in minutes - no research ops team required.
Details here: https://t.co/9RGsf374lw
Look forward to meeting many of you there!
The Prolific team will be at @LDNTechWeek 🇬🇧
Come speak to our experts at booth #352 about getting real human feedback into your AI workflows at speed.
#LTW#LondonTechWeek
Sharing our open-letter to the research community “Reply to Westwood: Questioning the empirical evidence that AI survey contamination is real and substantial” by David Rothschild, Soubhik Barari, Sunshine Hillygus, Trent Buskirk, and I.
[1/6]
https://t.co/43NUfWarqs
If you're at the @PsychScience (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
Dr @andrew_j_gordon presented "Not All Samples Are Created Equal" at #APS26BCN this week - the largest, most comprehensive comparison of data quality across online sample providers to date.
Critical work for researchers making platform decisions.
🧠 Avui arrenca l’APS Annual Convention 2026 al @CCIB_Forum. Del 28 al 30 de maig, el congrés internacional de ciència psicològica reunirà sessions plenàries, workshops, ponències, pòsters, flash talks, activitats de networking i sessions de desenvolupament professional per a investigadors i estudiants.
Enguany, el programa també incorpora sis Integrative Science Symposia (ISS) dedicats a explorar qüestions científiques complexes des de múltiples disciplines, amb temàtiques vinculades a la intel·ligència artificial, la salut mental, el desenvolupament cerebral, la cognició, el llenguatge, la cultura i les amenaces a la democràcia.
@PsychScience
-
@Fira_Barcelona
Autonomous agents are good at exploring within a frame, but not at stepping outside it.
Our team ran 50 autoresearch experiments on a DPO task - the agent flagged LoRA by experiment 11 and never followed up. Five minutes of guidance unlocked it 👇🏻
https://t.co/IJyGXedJMH
Two best-in-class tools. One continuous workflow.
Prolific is now natively integrated with Research Flow by @Typeform, combining our verified participant network with Research Flow's ability to uncover the "why" behind responses at scale.
The result = richer insights, faster than ever.
Explore for yourself here 👉https://t.co/fRmMiitY9U
Agents are everywhere in demos. But who's actually putting them into production?
Join AI Circle and Prolific for an in-person panel in SF next Tuesday, with guests from @GoogleDeepMind@GoogleResearch@Snowflake.
📅 Tuesday, May 26 · 5:30 PM
📍 San Francisco, CA
RSVP 👇
https://t.co/JBlXVhRGKI
AI misuse can affect research data in two ways. Prolific's authenticity checks catch both.
1) Bot authenticity checks (100% accuracy in testing) detect AI agents
2) LLM authenticity checks (98.7% precision) detect AI-generated responses.
Learn more → https://t.co/cdN8pJ7aEU
Prolific has been building human data infrastructure since 2014.
Over 10,000 organizations now use our platform, along with over 300,000 participants.
This recognition reflects that work, and the people who make it possible on both sides of our platform.
🚀🚀🚀
Prolific is in the Barclays Eagle Labs AI: 100, one of 100 companies to watch in 2026.
@eagle_labs@Beauhurst - thank you for including us.
Full report ➡ https://t.co/WuvqtPGSBb
Great co-hosting another panel with AI Circle in Seattle, with practitioners from @Microsoft, @Amazon, and @Braintrust. 🤝
The throughline → the industry's biggest bottleneck is still getting calibrated human judgment at scale. This is the problem we're solving.
At @AAPOR, Staff Researcher @andrew_j_gordon shares how Prolific approaches data quality. From 50+ verification checks, to lifetime account monitoring, to tools like our LLM and bot authenticity checks.
Great to have spoken with so many researchers about this at #AAPOR26.
@WalterO65851540 Hey @WalterO65851540 how soon a person gets invited off the waitlist is heavily based on supply and demand and so the wait time can differ for everyone - it could be hours for some and months for others. Please check your email regularly for an invite however!
Fraud in online samples is at an all-time high, and platform choice matters. This afternoon at @AAPOR, @andrew_j_gordon presents "Not All Samples Are Created Equal," the largest benchmarking study of data quality across providers.
🕓 4:00 PM
📍 Beaudry B, Lobby Floor
#AAPOR26
Senior Researcher @orioljbosch shares why @AAPOR is such an exciting moment for the field and for the Prolific team.
Coming up today: @andrew_j_gordon presents the most comprehensive comparison of online sample data quality to date. #AAPOR26