I was involved with evaluating IdentiFlight units at a wind project with a history of eagle kills. The majority of the 150 sites where IFI claims IdentiFlight is being used are provisional test deployments, and probably shouldn’t be considered happy customers just yet. If anything, word is getting around in some circles that this product isn’t ready for production use as many “customers” are quietly dropping their deployments. We are in the process of dumping ours.<p>We really need technology like IdentiFlight, because manual (human-induced) curtailment will not scale with the global growth of wind energy. But IFI is just a bit skeezy. It became apparent to me during our evaluation period that human operators were actually role-playing as the “AI” by live monitoring the camera feeds and taking actions manually to lend the impression that the system, once left fully to its own devices, would work sensibly and with more restraint than it actually does. IFI did not disclose this, and denied doing so when tactfully asked. But the difference in the system’s behaviour during and after the evaluation period is just too big a gap for me to ignore. I also became aware that the company generated and pushed for publication several papers in lesser peer-reviewed journals that claim to compare the efficacy of these units against human observers. The tests were, as far as I’ve seen, conducted under contract with environmental consulting groups eager to confirm the company’s desired conclusion using human observers with minimal experience and training, and the papers conveniently do not attempt to address the limitations of the IdentiFlight system in practice. And that brings us to this article, which is essentially more marketing, but masquerading this time as journalism instead of science.<p>I believe the Germans will be quite disappointed with these units once installed.