The State of the AI Ethicist.
Built entirely from gap report answers, rolled up, never individual. One person's sixteen answers tell you about them. A few thousand people's answers, aggregated properly, tell you something about the profession. That is the report we want to publish, once there are enough of you to make it safe.
Nothing to show yet. We need at least thirty completed gap reports before a single number here is safe to publish, see the count below.
Four rules, none of them optional.
The report is only worth publishing if taking part can never cost anyone anything. So the protections are structural, not promises.
Only the numbers
We collect the numeric answer to each of the sixteen gap report questions, nought to three. Nothing else about you rides along with it for this purpose.
Nothing ever joins it
If you ask us to email you your profile, that email lives in a completely separate store, never linked back to the anonymous answers used here. The two systems cannot be joined, by design, not just by policy.
The minimum-thirty rule
No cut of this data is published, by sector, seniority, region, anything, unless at least thirty people sit inside it. A group of five is a person with the serial numbers filed off. Thirty is the line, however interesting the number.
Aggregates only
Shares, medians, distributions. Never a table of individual scores, never a timestamp precise enough to match a single visit, never free text. Raw answers are processed into the aggregate and not kept longer than that takes.
Not enough respondents yet to publish anything.
We need at least thirty completed gap reports before the first real cut is safe to publish. When we get there, this page stops being a placeholder and starts being the report, with the count it is built on shown plainly.