Get 100% human participants, or twice your money back for any AI agent our powerful pre-study checks miss.

Overview

The 100% Human Guarantee supports high-quality research on Prolific.

If a submission in your study is identified as non-human through our Bot authenticity checks, you’ll receive double the costs for that submission as Prolific credits.

We’ve spent years building the infrastructure to verify every participant is human before they reach any study: including bank-grade ID verification, 50+ pre-screening checks, and real-time bot detection. We’re so confident in this system that we’re willing to back it up with this guarantee.

How it works

Every qualifying study is automatically covered by the 100% Human Guarantee.

To qualify:

Set up your study using Qualtrics and Prolific
Enable bot authenticity checks, which detect AI agents in your study.
We’ll credit you double the cost of any submissions flagged by our bot authenticity checks as low authenticity.
- This includes:
  - The participant reward
  - The platform fee for those submissions
  - VAT (if applicable)
Prolific credits are automatically added to your workspace balance within 21 days of the submission being flagged, and you’ll receive a confirmation email when they’ve been issued.

What credits will I receive?

We’ll credit you double the cost of any submissions flagged by our bot checks as low authenticity.

This includes:

The participant reward
The platform fee for those submissions
VAT (if applicable)

The credits will be added to your workspace balance within 21 days of the submission being flagged as low authenticity.

You can use these credits to replace the AI submissions by opening up more study places, or using them towards a future study on Prolific.

You don’t need to contact us to claim the credits, and you’ll be notified by email once added to your account.

Prolific credits are non-transferrable and non-exchangeable for cash or another alternative. Full details are outlined in the Terms and Conditions.

Which studies are eligible for the guarantee?

To be eligible for the 100% Human Guarantee:

Your study must be created using Qualtrics and recruit participants through Prolific
Bot authenticity checks must be enabled
Your study must be for legitimate academic or professional research purposes
Your study must be published after 31 March 2026, 2pm UTC and during the campaign period

All studies that meet these criteria during the campaign period are automatically covered, including pilots and duplicated studies.

What’s not covered by the guarantee

Studies that do not use Prolific’s Bot authenticity checks.
Studies using only LLM authenticity checks, as these are not designed to detect AI agents.
- LLM checks can be used along with Bot checks, but we won’t allow the results of LLM checks as proof that AI agents are in your study, as these are designed to detect when human participants are copy/pasting from tools like ChatGPT to answer free-text questions only.
Submissions not flagged as “low authenticity” by bot checks (this includes “mixed” or “high” results, or manually suspected AI use)
The 100% Human Guarantee campaign is not available to researchers based in Italy, Spain or Brazil due to legal or other regulatory reasons.
Studies must be run in good faith. Attempts to manipulate or artificially trigger low authenticity flags may result in removal from the guarantee.

Which studies should not include Bot authenticity checks and will not be covered under the guarantee?

Studies can ask participants to use AI tools for support (e.g. to look up information), but must not require the use of automated tools, scripts, or agents that complete the submission on the participant’s behalf.
Studies that require participants to use assistive technology also shouldn't use Bot authenticity checks, as this may result in participants being falsely flagged.

FAQs

How do Bot authenticity checks work?

Bot authenticity checks identify when an AI agent or AI-powered bot is answering your study, with 100% accuracy in testing. They look for non-human behaviors associated with automated environments across your whole set of questions, working with any question type.
We examine things like mouse tracking, pacing, whether it’s an automated environment answering, clicks, etc. They work by combining over 20 different checks to determine with confidence when it’s an AI agent vs a human answering your study.
We can’t be fully transparent about every check and threshold in our Bot authenticity checks, as that could enable bad actors to bypass them. However, if you’re uncomfortable using them in your study, you don’t need to include them, but your study will not be eligible for the guarantee.

Why can’t I use other survey software for the guarantee?

At this time, the guarantee only applies to studies run through the highly popular Prolific and Qualtrics integration. This is because Bot authenticity checks are only supported and verified with this integration.
We are working to expand support for authenticity checks across other platforms, as well as identify new high-accuracy methods for detecting AI agents.

Which methods does Prolific accept as proof that a participant wasn’t human?

Only studies that have Bot authenticity checks enabled are accepted to detect AI agents. Bot authenticity checks are the most reliable and validated method for detecting AI agents available for use with Prolific. Other methods are not covered under this guarantee because they don't meet the same standard of accuracy and often falsely flag participants. To maintain a fair platform for participants, we cannot accept other methods at this time.
We also do not accept Prolific’s LLM authenticity checks as part of this guarantee, as these are not designed to detect AI agents.

Can I run multiple studies under the guarantee?

Yes, all eligible studies run during the campaign period will be covered by the guarantee.

Can I still be credited under the guarantee even if I reject the “low authenticity” responses?

Absolutely! It’s up to your discretion whether you want to reject or approve “low authenticity” responses. Either way, you’ll receive twice the cost of the flagged submission(s) back in Prolific credits.

Is there a limit to how much credit I can receive under the guarantee?

There is no limit on the number of flagged submissions that can be credited under the guarantee Terms and Conditions.

I believe I'm entitled to credits, but haven't received them. What should I do?

If you haven’t received credits, check the following:

Does your study meets the eligibility criteria for the guarantee?
- Uses Qualtrics + Prolific together, with URL parameters selected as the method of integration
- Bot authenticity checks have been correctly set up like in the instructions here
- Saw results coming in from bot authenticity checks and some were flagged as “low authenticity”? (This would only happen if set up correctly)
If all set up correctly — did you receive a confirmation email to say you’ve been paid or not?
Have 21 days passed since the participant(s) were flagged as “low authenticity” by Prolific’s Bot authenticity checks?

If you’ve followed the steps above and still haven’t received credits, contact Support using the icon at the bottom right of this page.

Are submissions with “mixed authenticity” included in the guarantee?

No, only “low authenticity” results are covered under the guarantee.
“Mixed” results should be a very rare occurrence, but also mixed results mean we didn’t have enough signals to confirm high or low authenticity, and often can falsely pick up accessibility tablets or tools which we wouldn’t want to reject. We always advise reading the responses with submissions flagged as “mixed” and determining whether to reject or not based on that.

How do I opt out of the guarantee?

If you wish to opt out, click the text below to open the Chatbot window on the right side of your screen.

Opt out of promotion

You can still use Bot authenticity checks
If you tell us before you run your study, you won’t receive credits for any submissions flagged as “low authenticity” by our Bot authenticity checks
If you’ve already received compensation for flagged participants by the time you opt out, we can remove the issued credits from your account

Can I report suspected AI use that wasn’t flagged?

Yes, researchers can report suspicious or potentially fraudulent participants by submitting the relevant participant IDs through this form.

In the form, you'll provide details to help our data quality team investigate:

The data quality issue you've experienced
Participant ID(s)
Your study ID (optional)
Any additional data relating to the issue (optional)
If you’d like us to get in touch with the outcomes of your report, please enter your contact details (optional)
Note: Submissions that weren’t flagged by our bot authenticity checks aren’t eligible for credits under the guarantee.