(from the Algorithm Watch newsletter here https://r.algorithmwatch.org/nl3/dAy5-iYV4Kk4Edf6OpbOmg ) *All in for innovation.* Last July, I wrote a story about 200 undergraduates who failed their exams after a proctoring system had flagged their performance during their tests <https://algorithmwatch.org/en/spain-students-failed-blame-automated-proctori...>. I was in talks with a communication officer of the university in question. When he figured out that my reporting would not benefit the institution’s prestige, he tried to divert me from writing what happened and convince me of a supposedly much more uncontroversial subject: applying Artificial Intelligence to measure students' performances in the future. He did not say “emotion recognition” but described a software that could track students’ academic trajectories and scan their movements to assess, for instance, if their motivation might be dwindling. The university had been working on a similar project with IBM in 2018 <https://www.unir.net/actualidad/investigacion/inteligencia-artificial-para-m...>, but at the time, it didn’t involve real-time image analysis. The software is yet to be deployed. With all the marketing effort to hype the technology, one can guess that the project might have been dropped (or delayed) because the implementation proved to be more difficult. *Tech vs. psychology.* It’s not the first time (nor will it be the last) that education institutions toy with face and emotion recognition systems. Some secondary schools in France were forced to ban face recognition <https://etudiant.lefigaro.fr/article/la-justice-interdit-la-reconnaissance-f...> on the premises. Teachers in Denmark use apps to measure the “mood landscape” <https://www.technologyreview.com/2023/04/17/1071137/denmark-teachers-apps-st...> of elementary students. The apps have costed the government millions of crowns. In Spain, almost every autonomous region has incorporated these tools in their Artificial Intelligence strategies. They plan to improve the academic careers of students, either by scanning grade histories or measuring their behavior in real time. None of them have actually moved forward with it <https://www.businessinsider.es/estrategias-ia-espana-no-estan-sirviendo-nada...>, but it’s /cool/ and avant-garde to say they will. They are not completely to blame for trying, taking into account that even the Organization for Economic Co-operation and Development (OECD) is pushing AI to “assess knowledge” and “gauge emotion” <https://archive.ph/I9Nb9>. *Outdated ideas.* The people heralding these tools as the future of education seem to care little about researchers criticizing the pseudo-scientific approach of emotion recognition technology <https://www.article19.org/emotion-recognition-technology-report/>. The use of AI to predict people’s behavior or interpret their emotional state is fundamentally based in the assumption that the expression of emotions is universal and limited to six basic states <https://www.annualreviews.org/doi/abs/10.1146/annurev.ps.30.020179.002523>. This is scientifically untenable and has been harshly refuted <https://journals.sagepub.com/doi/10.1177/1529100619832930>. A smile doesn't always mean happiness, just as a wide range of expressions can indicate anger. Even if automated systems can detect an expression, they don't know what to make of it. Yet, some enthusiasts trust that AI will not only swiftly interpret human emotions but also predict people’s behavior and make decisions on that basis. Minors and migrants (or both) are a testing ground <https://pubmed.ncbi.nlm.nih.gov/37535569/> in many cases. The recently-approved AI Act bans emotion recognition systems in schools but carves out exemptions for “medical” or “safety” reasons. It remains to be seen if the AI Act will protect students and pupils from transgressive surveillance rooted in pseudo-scientific discourse. -- -- EN https://www.hoepli.it/libro/la-rivoluzione-informatica/9788896069516.html ====================================================== Prof. Enrico Nardelli Past President di "Informatics Europe" Direttore del Laboratorio Nazionale "Informatica e Scuola" del CINI Dipartimento di Matematica - Università di Roma "Tor Vergata" Via della Ricerca Scientifica snc - 00133 Roma home page: https://www.mat.uniroma2.it/~nardelli blog: https://link-and-think.blogspot.it/ tel: +39 06 7259.4204 fax: +39 06 7259.4699 mobile: +39 335 590.2331 e-mail: nardelli@mat.uniroma2.it online meeting: https://blue.meet.garr.it/b/enr-y7f-t0q-ont ====================================================== --
