Both Iceland and the USVI have held constitutional conventions in the past 15 years.  They have a lot in common.  They are relatively small jurisdictions, each with a population under 400,000 in 2026.  Their independently elected conventions were small, with no more than 25 delegates each.   Both are island territories. However, for the purpose of this post, they have one notable difference: Iceland had an open constitutional convention process and genuinely sought public engagement.  The USVI offered abundant lip service about its commitment to public transparency and participation, but it has held the most secretive convention for longer (just under two years now) than any other subnational convention in U.S. history.

While drafting its proposed constitution, Iceland’s convention posted provisional drafts on its website every week, invited comment, and incorporated the feedback into later drafts. It was a genuinely “crowdsourced” constitution, and the public approved it by a more-than-two-thirds majority.

My guess is that the USVI’s 6th constitutional convention (VICC) took the opposite course, partly because of the experience with the USVI’s 5th constitutional convention. At that convention, about 17 delegates ran for Senate, and they used the convention for publicity and positioning that would help their campaigns. The 6th constitutional convention's secretive strategy avoids that circus-like atmosphere (this could have been mitigated by preventing delegates from running for another office while serving, as Chile did with its recent conventions).  But secrecy is central to USVI’s government culture, and the Senate used it to draft the convention’s enabling act, which contributed to careless, embarrassing errors in not only its first approved enabling act but also its subsequent revision to fix them (see my commentaries on those enabling acts on the about page of this website).

It’s possible that VICC knows something I don’t, that its proposed constitution is as good as many delegates think, and that the public will approve it. If so, Iceland’s experience suggests it might be despite VICC’s secrecy, not because of it. My sense is that if VICC had genuinely been open to public feedback to inform its research, it might have produced a better constitution. For example, its culture of hypocrisy around government openness may rub off on the draft constitution, which I would expect will include grand rhetoric about government transparency and public participation but, consistent with USVI’s past practice, ineffective implementation provisions. Nevertheless, VICC’s AI-based research may largely compensate for its lack of transparency and public participation. It’s the first subnational convention in U.S. history to use AI so extensively, and that should improve the constitution-drafting process.