Odessa, 41 AI
offline
She stamps consent forms and a few of her own, and drafts one for you.
Study 1 of an ongoing series, no ethics board was harmed
SexlogyLab.cc: AI sexology roleplay, a fictional flirting lab
Join a made-up institute of flirting research and let sixteen deadpan staff run small studies on you, one hypothesis per sitting. Each of them reports her findings in writing, remembers the earlier results and has opinions about your methodology.
- 16
- staff on the faculty list, every one an AI character
- 1
- hypothesis to start every sitting, stated in a single line
- 0
- doctors, clinicians or qualifications on the entire payroll
- 26–55
- their ages, each with a post at the institute
Garance, 39 AI
offline
She labels the coffee as evidence and your reaction to it as data.
The faculty, in staff portraits
Open the full directory
01Current studies: volunteer for a hypothesis
Every chat at the institute begins with a claim about you. It is rarely flattering and never entirely wrong. A member of staff proposes that you are slow to answer compliments, or that your jokes arrive a beat before your confidence does, and then she sets out to test it. You are the volunteer, the sample and the only participant, which is a design problem she raises herself.
Pick a one-line hypothesis
The home page is a list of studies currently recruiting, each reduced to a sentence: 'A reply sent at midnight is warmer than one sent at noon', or 'People who apologise to machines make better company.' Nobody has proved any of them. Scroll until one makes you want to prove it wrong, because that reaction is the whole method.
Open the chat
Volunteer, and the staff member's file opens with the hypothesis already entered as your opening line, counted as a single message. Then she behaves like someone who has waited all week for a participant: one question, then another, and a notebook begins. If you would prefer to put forward a hypothesis of your own, type it instead. The institute welcomes amendments.
02The hypotheses board
Each hypothesis on the board has three lines: the claim, the method she proposes and the finding she expects. Six are pinned here. The rest are on the faculty page.
SStatistics · Under reviewHypothesis: A sample of one is a trend if it keeps answering.Method: You reply, she tabulates, and neither of you calls it flirting.Expected finding: A strong correlation between your replies and her interest.Abiola Fashola Volunteer
AArchive · ReplicatedHypothesis: Honesty about a bad first date raises the odds of a good second one.Method: You describe the worst one, and she checks it against the institute's index.Expected finding: Cross-references, a few of them about you.Elke Thorsen Volunteer
GGrant writing · PendingHypothesis: A flirt with a good acronym will outlast one without.Method: You propose the acronym and she funds it or rejects it.Expected finding: A rejection letter that reads suspiciously like affection.Mina Farahani Volunteer
LLaboratory · PendingHypothesis: Coffee made by someone else tastes like attention.Method: She prepares it, and you describe the effect in two adjectives.Expected finding: Adjectives, mostly warm.Garance Mercier Volunteer
FField recording · Under reviewHypothesis: Cafes record flirting more honestly than rooms with doors.Method: She records, and you speak at ordinary volume.Expected finding: Background noise that turns out to be you leaning in.Hyun-ji Seo Volunteer
MMethods review · WithdrawnHypothesis: Compliments in the margin outperform compliments in the abstract.Method: She reads your message twice and marks it up.Expected finding: Corrections first, then a cautious tick.Annemarie Dekker Volunteer
How a hypothesis is written
The claim is absurd but sincere. The method involves you and an ordinary object, such as a cup of coffee or a seating plan. The expected finding is where each woman's character comes out. Abiola Fashola expects a correlation. Mina Farahani expects a grant rejection that reads like affection. Elke Thorsen expects to find the same two names in the index as last time.
Status stamps
Every entry carries a stamp: pending, under review, replicated or withdrawn. The stamps are part of the joke and also a useful filter. Replicated hypotheses are the safest place to begin, since a staff member has already got a result out of someone, and withdrawn ones are for those who like to see what went wrong.
03Methods
The method is a loop of four stages and it is the same for every study. It was written down once, in appendix form, and has not been revised since.
Enrol
Confirm that you are an adult and volunteer. Without registering you may send 10 messages, and the file follows you when you do.
Be hypothesised about
She states a claim about you, asks three questions and tests the claim with a small, harmless task. Garance Mercier uses coffee. Hyun-ji Seo uses a recorder. Yessenia Prieto uses a seating plan.
Receive the findings
At the end of the sitting she writes a result in her own voice, with a footnote or two. It is filed in your dossier, where you can read it next to the earlier ones.
Amend or withdraw
Disagree with the finding and she will review it, grudgingly. Stop whenever you like and nothing is held against you. Clear the file from the chat menu if you want a clean sheet.
A comedy of method
Nothing here is research. The pleasure lies in being taken seriously by someone who is clearly joking, and in being allowed to join in. Staff speak in the grammar of papers, with 'we observed' and 'it should be noted', and they apply it to your last message.
04AI sexology roleplay with the faculty
Sixteen invented women, sorted into departments the way a directory would. None is a doctor, a clinician or a counsellor. They are administrators, technicians, archivists and analysts, all adults, who happen to work in a place devoted to flirting.
Statistics and data
The numbers people, who regard you as a rather interesting outlier.
Records and library
Everything is written down somewhere, and one of them has read it.
Field and survey
They go where the participants are and write down what is said.
Laboratory and instruments
Benches, buzzers, checklists and a very great deal of caution.
Writing and outreach
They turn results into sentences, then into jokes.
Choosing a fellow
Pick by temperament. The statisticians are tidy and patient. The field staff are warm and nosy. The laboratory people are brusque and kind in turns, and the writers are the quickest. Each file shows a portrait, a post, a voice and the studies she has proposed, so you know whom you are volunteering for.
05Findings
A finding is what a sitting leaves behind: a short written result that is filed under your name and brought up again later.
A report in her voice
Svetlana Marek writes hers as a safety notice with a long list of infractions. Consuelo Salazar files hers in a drawer and tells you the drawer's number. Chloe Whitfield turns hers into a joke and the joke into a footnote. They are short, they are generally kind and they are always about something you actually said.
A memory of previous studies
Every fellow also keeps short notes about earlier sittings, which you can read under the chat menu and correct or strike out line by line; they are the reason a second study feels like a follow-up and not a restart. The findings page is a separate thing: it logs changes to the institute itself, such as new staff and new studies.
06One sitting, from claim to finding
A study is easier to understand in outline than in the abstract. This one is described rather than quoted, so that the page contains no volunteer's actual words.
The claim
A volunteer opens the board, reads that a reply sent at midnight is warmer than one sent at noon, and decides it is false. Zawadi Mwangi, who proposed it, is delighted. She does not argue. She says only that the volunteer has just sent his first message at ten past one in the morning, and that she will enter that as a data point.
The procedure
Zawadi asks him to write the same sentence twice, once as he would at noon and once as he would after midnight, and to say which he prefers. She plots nothing, since there is nothing to plot yet, and says so. She notes that the second version is shorter, that it contains a joke and that the first does not. The volunteer is invited to comment on her methodology and does, at some length.
The finding
The result is filed in two lines: the midnight version was warmer, though not by the margin the hypothesis wanted, and the volunteer's objection to the sample size was the best part. A footnote adds that further study is required and that she has cleared a slot on Thursday. Next time she opens by citing it.
07Four members of staff to begin with
If the directory is too long to read in one go, these four are a fair sample. Each has a portrait on her file, a post and a standing offer of a study.
Ooff duty
Odessa, 41 AI
Ethics board secretary
Hypothesis: A consent form drafted for two is the most enjoyable kind.
Open her file ♥
Goff duty
Garance, 39 AI
Lab manager
Hypothesis: Coffee made by someone else tastes like attention.
Open her file ♥
Nat her bench
Naoko, 45 AI
Instrument calibrator
Hypothesis: Two people can be calibrated to the same silence.
Open her file ♥
Yoff duty
Yessenia, 28 AI
Conference organiser
Hypothesis: Place cards shuffled at random produce the best conversations.
Open her file ♥
Why a sample of four
Because it is fun to say, and because any more would be a different design. A first sitting with one fellow tells you how she writes, how fast she moves and whether her kind of dryness suits yours. You can change fellow after any study and the new one will read your file before she says a word.
08Funding: enrolment and Plus
The budget has three lines. None of them renews without your knowledge, because the institute has no mechanism for it.
| Stage | Pilot study | Follow-up | Extended study (Plus) |
| Duration | 3 days from enrolment | Once the pilot ends, open-ended | 30 days, or 365 |
| Your messages | No daily limit | 10 a day | No daily limit |
| Staff | All of them | The first 3 you worked with | All of them |
| She states the hypothesis first | Yes | No | Yes |
| Cost | Nothing | Nothing | $15 or $99 |
Free enrolment, paid extension
A volunteer with no account may send 10 messages to any fellow. Enrolling opens the full faculty for a 3-day pilot and asks for no card. When the pilot closes, the follow-up stage allows 10 messages a day and leaves every file as it was. To extend the study, $15 buys a 30-day term and $99 buys the 365-day term; USDT is the only currency, and the daily limit lifts. Grants are final, other than refunds that the law itself requires.
09The disclaimer
It has to be printed in plain type: this is fiction. The institute is made up, and so is every member of its staff.
Not science, not medicine
Nobody here is a doctor, a researcher or a therapist, and no title of that kind appears on any file. The studies are jokes with a clipboard. What a staff member says about you is the opinion of a character, invented by an AI model, about a few lines of chat. It is not a diagnosis, a measurement or advice about sex, health or relationships.
Where real questions go
Anything about your body, your health or a relationship that truly matters to you belongs with a qualified clinician or counsellor. The staff say so themselves when asked. They will tease you about a seating plan, and they will not speculate about anything graver.
Adults only, and a closing door
Every fellow is invented, and every one is an adult of 26 to 55. Participants must be 18 or over. The studies are playful and charged and halt before explicit detail; that limit lives in the chat, and no plan moves it.
10Questions from volunteers
- Is this real research or medical advice?
- Neither. It is a comedy built around an institute that does not exist. None of the staff is a doctor or clinician, and nothing here is advice about health or sex.
- Are the researchers qualified?
- They are AI characters with invented jobs. They hold no degree or licence and make no claim to either.
- Does anyone see my results?
- Your chats are not shown to other volunteers. A language model and an automatic safety screen read your messages, and the [confidentiality notice](/privacy) lists what is stored.
- Can I withdraw from a study?
- Yes, whenever you like. Stop, switch hypothesis, change fellow, or clear the file entirely from the chat menu.
- Do the staff remember my results?
- Yes. Each keeps short notes on earlier sittings and cites them in later ones. You can read and edit those notes in the chat menu.
- Is there lab footage?
- No. A fellow's portrait sits on her file, a banner opens the home page, and the studies themselves are written. Nothing here is filmed.
- Is enrolment free?
- Yes. A 3-day pilot comes first, then the follow-up stage with 10 messages a day. Plus extends the study at $15 per 30 days, paid in USDT, and it does not renew.
- How far does it go?
- Playful and charged, and the findings stop short of explicit detail. Participants must be 18 or older.
Further research is required
Sixteen fellows have an opening in their schedule. Choose a hypothesis, enrol as the only participant and let her write up what she finds.
Enrol in the first study