A Brief — 2026-06-16
Early AI adoption in research used to be called front-running: strategic, forward-looking, a competitive advantage. That framing assumed the terrain was open — that moving early was simply moving smart.
It is no longer open. Platforms gate access. Institutions set policies. Funding bodies attach conditions. The same behaviour — adopting AI tools before institutional sanction, collecting API keys, building outside official channels — is now structurally trespassing. Not illegal, but transgressive. Moving through terrain that wasn’t designed for you, before or despite the gates.
Herumlungern im Gelände. Loitering on the land.
| Front-running | Trespassing | |
|---|---|---|
| Terrain | Open — no gates yet | Closed — gates exist or are being erected |
| Action | Moving early on information others don’t have yet | Moving through space not designated for you |
| Origin | Financial markets: trading ahead of public information (illegal) | Property law: entering land without permission (civil, not criminal) |
| In research | Early AI adoption before institutional guidance exists | Same adoption, once policies define what is sanctioned |
| Framing of actor | Strategic, forward-looking, individually advantaged | Transgressive, unsanctioned, institutionally problematic |
| Who can do it | Those with information access and nerve | Those with time, confidence, and connectivity — unevenly distributed |
| Policy response | None needed — the system benefits from the pioneers | Warning signs and guards; managed sandboxes; bounded environments |
| Structural implication | Informational asymmetry as individual advantage | Spatial asymmetry as collective inequality |
The shift from one frame to the other was not caused by a change in behaviour. It was caused by a change in terrain — institutions closing the land that was previously open.
Short-term contracts are structurally mismatched with current innovation speed in engineering disciplines. A two-year postdoc, a one-year project extension, a six-month teaching contract — none of these timescales are compatible with a field where the tool landscape changes quarterly. Researchers on precarious contracts cannot afford to wait for institutional guidance that will arrive after the relevant window has closed.
The gap between institutional response time and innovation speed is not closing. It is widening. As it widens, more researchers will trespass — not because they are reckless, but because it is the only way to stay competent. The trespassing trend will rise.
Trespassing is generative. It produces knowledge, contacts, and prototypes that institutional pathways would not have sanctioned in time. The trespassers are doing necessary work — mapping the terrain, finding what is useful, building proof-of-concepts that eventually inform policy.
But trespassing is not equally accessible. It requires time, confidence, and connectivity — resources distributed unevenly across the research workforce. Early-career researchers on short-term contracts are both the most pressured to trespass and the least protected when it goes wrong.
The current institutional response — warning signs and guards, managed AI environments, bounded sandboxes — does not resolve this. It just makes trespassing more legible as transgression.
The answer is not to guard the terrain more carefully, and it is not to celebrate trespassers while leaving the structural conditions unchanged.
It is to build commons infrastructure that makes the trespassing move unnecessary — where what the early adopters were sneaking access to becomes genuinely public. Compute is the clearest case: publicly funded compute, governed as a commons rather than a centralized subsidy, removes the API key bottleneck that currently gates participation.
This reframes the policy question. The choice is not between controlled AI access and uncontrolled AI adoption. It is between infrastructure that reproduces inequality (gate-kept, institution-mediated) and infrastructure that distributes access by design (commons-governed, collectively bargained).
Funding the commons is the structural fix that makes trespassing obsolete.