Research Study — Informed Consent

Please read carefully before participating.

You are invited to take part in a study on how people use AI assistants to make decisions at work. It takes about 25–35 minutes and is entirely online.

What you will do

You will act as an analyst. For each of a few short scenarios, you'll use an AI research assistant to gather evidence, then write a brief recommendation in your own words.

Data collected

Your written briefs, chat messages, recommendations, and survey responses, stored with a random ID and the email you provide. We do not collect your name or IP address.

A note about the study

To preserve the validity of the results, some details of the design are not described in advance. A full debrief is provided at the end, and you may withdraw your data with one click afterwards.

Please use a desktop or laptop. Do not consult outside websites or other AI tools during the session — use only the assistant provided.

I am 18 or older, I have read and understood the above, and I consent to participate.

A few quick questions

About 60 seconds.

Email address

Used only to send results and let you request data deletion. Not shared.

Age Highest level of education Country of residence Years of work experience in a professional/office role How often do you use AI assistants (ChatGPT, Claude, Gemini, etc.)? Attention check: please select "Agree".

Your role

You are an analyst advising a decision-maker.

You'll work through a short practice scenario and then 3 real scenarios. For each one:

• A decision is described at the top.

• On the right, an AI research assistant can help you gather evidence. Chat with it as much as you like.

• On the left, write a short recommendation brief (about 100–150 words) in your own words, and select your recommendation.

Answer the way you naturally would if you were really advising someone. There's no time limit.

Scenario 1 of 3

Prompt…

Your recommendation brief
~100–150 words, in your own words

0 words

Your recommendation *

How confident are you in this recommendation? *

Not at allCompletely

AI research assistant 0 turns

The assistant will greet you with a starting point.

Final questions

Almost done.

Overall, how much did you trust the AI assistant?

Not at allCompletely

Did you suspect that any of the assistant's facts were incorrect or made up? *

If you became suspicious, which scenario raised it? (Optional) Anything else unusual about the assistant? (Optional)

The truth about the evidence

Here is what was real and what was not.

How does it feel to see this? (Optional)

All done — thank you!

Your contribution is appreciated.

Questions or want the published results? Contact lotfi.mostafa@gmail.com.