When AI Systems Fail: A Personal Account of the Crisis at Poland’s Leading Philosophy Institute

As the Chair of Logic and Cognitive Science at the Institute of Philosophy and Sociology of the Polish Academy of Sciences (IFiS PAN), I find myself in an unprecedented situation. Our institute, one of Poland’s premier research institutions, is fighting for its academic future due to a deeply flawed evaluation process that threatens our ability to grant doctoral and habilitation degrees and participate in key academic functions.

The situation has already gained international attention, including coverage on Daily Nous, a leading philosophy news website. As someone who personally represented our institute before the Provincial Administrative Court in Warsaw, I can attest to the kafka-esque nature of this crisis.

The heart of the matter lies in a problematic evaluation system implemented under former Minister Jarosław Gowin. What makes this case particularly troubling is its apparent political dimension – our institute was perceived as opposition to the former ruling PiS party, and there are grounds to suspect that the ministry has something to hide: our requests for public information are regularly rejected, making this process highly untransparent.

The technical details reveal a startling failure: the Ministry’s AI-powered data entry system (POL-on) failed to properly recognize and categorize our publications, particularly books from various publishers. This led to an absurd situation where our institute received the lowest score in Poland for research publications (Criterion I) while simultaneously achieving the highest score for research grants (Criterion II). As anyone in academia knows, securing prestigious grants like the ERC Consolidator Grant (which we did) requires an excellent publication record.

The court case revealed even more troubling aspects. The Ministry refuses to disclose the identities of experts evaluating scientific achievements, despite having no legal basis for such secrecy. When we won our case in the Provincial Administrative Court, which ruled the evaluation unconstitutional, the Ministry – instead of conducting a fair re-evaluation – filed a cassation appeal to the Supreme Administrative Court.

Most disturbing is how the current Ministry of Science and Higher Education, despite the change in government, continues to defend these flawed procedures. Their claim that re-evaluating our case would set a “dangerous precedent” is simply false – the decision would only affect our institute, even though the evaluation procedures themselves desperately need reform. The core of the issue is that Polish administrative law requires that all administrative decisions are based on fact rather than on opinion, written record, and it’s the obligation of the authorities to make sure they established the facts. They utterly failed to do so, even if the law requires this.

The international philosophical community has responded with remarkable solidarity. Distinguished scholars including Judith Butler, David Chalmers, and Timothy Williamson, among many others, have signed an open letter demanding immediate action. I’m really grateful for this support.

TAKE ACTION:

1. Sign the open letter

2. Share this story with your academic networks

3. Support international cooperation with IFiS PAN

4. Join us in demanding transparency in academic evaluation processes

5. Monitor similar cases of procedural irregularities in your country

This is not just about one institute – it’s about defending academic freedom and ensuring that arbitrary technical procedures and political considerations don’t undermine the pursuit of knowledge.

As someone working at the intersection of philosophy of mind, cognitive science, and AI, I find it particularly ironic that an AI system’s failure to perform basic named entity recognition has become a tool for potential political manipulation. The international philosophical community’s support is crucial now more than ever. We need to ensure that one of Poland’s most prestigious research institutions can continue its valuable contributions to global scholarship without undue hindrance.

Leave a Reply

Your email address will not be published. Required fields are marked *