Skip to content
AI AgriBench

Governance

Structure

A Steering Committee of academic, industry and nonprofit representatives governs the consortium, with working committees reporting to it. The Center for Digital Agriculture at the University of Illinois Urbana-Champaign leads it. Decisions follow majority consensus with Steering Committee oversight.

Neutrality rules already in force

  • Member companies' scores are reported exactly as non-member companies' scores are.
  • No model is ever used to judge itself.
  • Marketing material that singles out a member's performance is not published by the consortium.
  • Member quotes are attributed to the individual and the organisation, and are never presented as consortium positions.
  • The leaderboard is public and free. Results are not gated behind a paywall or a member-only portal.

Funding disclosure

Funders to date: AIFARMS, the Center for Digital Agriculture, and the Gates Foundation. Several consortium members build or sell AI advisory products that the benchmark evaluates. That is disclosed here rather than left to be discovered.

Conflict of interest

TODO (Committee): a written policy covering — who may be involved in scoring a system their employer builds; recusal rules for Steering Committee members whose organisations appear on the leaderboard; how membership fees are separated from evaluation decisions; and what a member may and may not claim about its own results.

Corrections

TODO (Committee): the corrections policy. The principle the site is built on is that an archived release is never edited after publication; a correction is issued as a new, dated post that links back to the release and states what changed and why. The committee needs to confirm who authorises a correction and within what timeframe.

Crisis communications

TODO (Committee): what AI AgriBench says if a member company scores poorly and reacts publicly, and who responds if the methodology is challenged in peer review.

Data governance

TODO (Research team): how contributed farmer–expert data is licensed, anonymised and retained, and what a contributing organisation may withdraw.