IC AI - Prolific Researcher Reviews

Unlike some of their other video study descriptions that reveal the long and arduous screening and setup process, this study description only says: Before the rating starts we will ensure that your setup is good for detecting slight differences in quality. We will guide you to make sure all settings are as expected. After reading that, I thought, "Good, their 12-minute intended time must be right, and I won't have to go through all of that ridiculous stuff." I was wrong. The setup was worse than with their other video evaluation studies. This study took me 22 minutes instead of the estimated 12. I know that they purposely do this because after the initial setup they go faster, but IC AI, in their infinite wisdom, has decreased their already annoying recalibration from 45 minutes down to 30 while lowering the pay from $3 down to $2. They have also removed any potential bonuses and doubled the number of videos.

They asked me to return two submissions due to supposedly missing attention checks. It is impossible to miss attention checks because you can't even answer the questions unless you play the recordings on a loop. These "researchers" are fraudulent.

Scam. They claim you failed attention checks so they don't have to pay you.

After doing 130 of these across two back-to-back projects, I can safely say there are some logistical issues. 1. Calibrations: I think these are a huge pain. The first one is the worst. The recurring 45-minute recalibrations are tedious and, in my opinion, unnecessary. I can't expect most monitors to have gamma and color drift that bad. Still, as long as you are mentally prepared for the recalibrations, you should be okay. In my first batch of 80, I had to return 4 because I failed the recalibration. This is where there's another issue. One of the tests is an image test. The question is, "Which is the better image?" A simple question, but if you don't pass, or even if you do pass, you might be greeted with conflicting information. This is what threw me off. At first, I was told I might be too close to my monitor. This was with a passing score, but when I failed, it said I might be too far. The thing is, I was the same distance both times. The only time the system doesn't give you this kind of feedback is when you get a perfect score. 2. Another issue is the hardware detection. In the first 80, I received an error message telling me my browser wasn't compatible about 4-5 times. There's another review here with the specific message. However, in the current iteration, I have done 50 tests and have gotten this message about 6-8 times. It's an annoying message because you have to refresh to do the test again. If you're too close to the 45-minute mark, you might trigger an unexpected recalibration. 3. Feedback: While I don't typically look beyond an approval as feedback, it's nerve-wracking to perform 80 studies wondering if you did them well. That could just be me. Nevertheless, the approval time wasn't bad. It was only seven days from the start of the first project. I was approved for all 76 of the studies I did. The bonus was nice too. I also appreciated the communication with the researcher. I provided feedback, and they responded rather quickly. When the second project appeared and I asked them about participating in it, they responded quickly again. However, they didn't address my anxiety regarding my pending studies. Overall, I don't think this is a bad study. There are control clips and attention checks. It's not an easy study. It could pay more for the amount of effort I put in, but I'm thankful I had the opportunity to participate. September 2, 2026 update Originally, I gave this study 4.5 stars, but I have dropped that to 3. Firstly, because I participated in the second round of 80 studies, and it took them almost three weeks to review and approve my 80 submissions. During that time, they launched another project, which put me out a bit. If they can launch more projects, they should review those who have already worked for them. It's one thing to wait for a single study to be approved, but when you're waiting for 80, that is completely different. While Schrödinger's cat is an interesting thought experiment, Schrödinger's money is completely different. The reason for the other drop (or half-star drop) is technical issues, calibrations, etc. When the project task is comparing what is supposed to be the same video, and yet they aren't the same, that is confusing. Coupled with the constant 45-minute recalibrations and browser errors, it's a pain. All in all, I would hesitate to engage too heavily with these folks unless you don't mind waiting weeks to get paid.

DO NOT USE THIS RESEARCHER. I returned this. The reason is the time it would take to finish this study is not even close to the time they have shown as the actual completion time. The time shown for intended time is 25 minutes. The actual time is shown as 24 minutes, There is no way that time is even possible. I was probably only about halfway through this study and was already beyond 25 minutes. To complete this study would easily take another 15 to 20 minutes. Meaning the full ACTUAL completion time would have been around 40 to 45 minutes, if not MORE. It would have timed out. As it is I gave them a free 25 minutes of my time. No more.

I made 60 submissions and had to return the study 7 times due to issues during the screening process. When I opened the study page, after a minute of loading, about half of the time the site would complain that my browser couldn't play H.264 4:4:4 content, even though main study videos (that i'm pretty sure were H.264 4:4:4 content) played just fine. Refreshing the page until the warning went away solved the problem. I was able to pass the rest of the screening steps every time they appeared. Around submission 30 the site kicked me out for not having a sufficient internet connection, threatening rejections if I were to ignore the warning. Refreshing the page would give me the same error, and then refreshing again gave me a message saying something along the lines of "we don't have any tasks available for your device/connection. please try some of our other HITs" so I left the study alone for an hour or two. I'm guessing that had something to do with site traffic. When I opened the study page again it made me go through the entire screening process again, and then I was able to make more submissions. A few hours after the study concluded, despite all the issues, every one of my 60 successful submissions were approved.

Waiting until Prolific bans them. Remember them from Amazon Turk years ago, where after attempting over many years to qualify, I never could despite having a high speed wi-fi connection and a Macbook. I didn't even get the .50 for this one , the screenout was .14. But I took the assessment. I wasn't *screenedout*- I was disqualified, technically, but after I fully did the same amount of work as someone who would have qualified. The last batch they had in here wouldn't even let anyone use a MacOS, which is ridiculous. And prior that, they had a batch which caused me anxiety knowing they were going to reject something but I did a bunch anyway. I was asked to return three of the bunch(maybe 20) being told "they were obvious answers I got wrong". Uh uh. I know for a fact I answered correctly based on the guidelines. Why of the 1,000 or so surveys Ive done on Prolific have I never been asked to return a study for poor results, yet IC AI had me return three in one day? Curious. They are rated very poorly for a reason.

Low pay for AI training, unpaid screen-out, and was told to return the study.

Unclear instructions, poor communication. Won't respond to questions/messages. Bars you from completing more tasks without giving clear guidance as to what was wrong with previous submissions.

The calibration section before the study begins is a nightmare.

Again, I answered all the questions, it said I had passed the test, then it asked me to "check my answers again" and wouldn't let me move on. When I changed one answer, I got screened out with no pay and a threatening message.

There was some setup followed by some qualification testing that was not easy, then it immediately screened me out with no pay. The screen-out page reads: "There is no assignments that match to your profile now. Please try it again in two-weeks time. We thank you for your participation. NOTE: Any try to perform assignments and neglecting this message is forbidden and leads ro rejection of all of your submissions." (Typos and bad grammar are verbatim.)

Would not load

Timed out and then rejected after completion. The set up took me like 30 minutes.

Calibration process is near impossible, no matter what I set my screen brightness to, it tells me I missed a shape or higher quality image and screens out near every time

They seem to approve, but they suck to do these tasks. It's specifically the calibration and the rate at which they make you calibrate which sucks..

Really surprises me about the rejections, I know the qualifier screeners give people trouble but I had 1 return requested out of I don't even know how many. They also have started approving pretty quickly used to be close to the full 22 days.

Rejected after completion and won’t give option to return.

The screening process is insane and good luck passing. Same for the video ones.

Rejected after completion and won’t give option to return