About the project, the study and the funding

It is an independent research project studying what counts as human capability now that machines can write, analyze, code and create — and what evidence of that capability organizations should trust when they make decisions. The 2027 study is the first wave of what we intend to be annual research.

Employers, educators and institutions still have to judge what a person can do. The tools those judgements rely on were designed for a world where the artifact and the ability were the same thing. There is very little evidence about what works now, and decisions are being made anyway.

The Human Evidence Project was founded and is funded by SignalVerified, a company working on provenance and evidence of contribution. SignalVerified has a commercial interest in this subject, which is why the funding is disclosed on every page rather than in a footnote.

Research questions, the instrument and the analysis plan are written and published before data collection begins. The anonymised dataset and full methodology are released with each publication. Findings are published in full, including any that are unhelpful to our funder, and the funder does not review results before release.

It is the project's first annual study: a single wave of structured research with professionals and decision-makers across knowledge-work sectors, fielded in 2027, examining how capability is demonstrated, what is accepted as proof of contribution, and how decisions change when the origin of work is uncertain.

The study is aimed at people who produce or assess professional work — for example those who hire, review, teach, commission or manage it. When participation opens, the eligibility criteria and screening questions will be published with the instrument.

About fifteen minutes. The instrument combines short answers about your own work with a small number of judgement tasks.

Responses are collected anonymously and analysed only in aggregate. We do not ask for your name, and we do not publish anything that could identify an individual or a specific employer. Anonymised data is released alongside the findings so the analysis can be checked.

The report, the anonymised dataset and the full methodology are published together after the 2027 wave closes. They are released publicly at the same time, without an embargoed or privileged version for the funder.

We welcome review of the instrument and the analysis plan before fielding, as well as replication and secondary analysis afterwards. Write to the research team and we will share the current draft materials.