My new paper, 𝐓𝐡𝐞 𝐇𝐢𝐝𝐝𝐞𝐧 𝐒𝐜𝐚𝐥𝐞 𝐨𝐟 𝐄𝐦𝐨𝐭𝐢𝐨𝐧 𝐚𝐧𝐝 𝐅𝐫𝐞𝐞 𝐖𝐢𝐥𝐥: 𝐍𝐞𝐮𝐫𝐨𝐬𝐜𝐢𝐞𝐧𝐭𝐢𝐟𝐢𝐜 𝐚𝐧𝐝 𝐂𝐨𝐦𝐩𝐮𝐭𝐚𝐭𝐢𝐨𝐧𝐚𝐥 𝐏𝐞𝐫𝐬𝐩𝐞𝐜𝐭𝐢𝐯𝐞𝐬, published with 𝐁𝐞𝐚𝐭𝐫𝐢𝐜𝐞 𝐒𝐞𝐜𝐜𝐨𝐦𝐚𝐧𝐝𝐢 and 𝐅𝐢𝐥𝐢𝐩𝐩𝐨 𝐒𝐚𝐧𝐟𝐢𝐥𝐢𝐩𝐩𝐨, is now online on 𝐈𝐄𝐄𝐄 𝐗𝐩𝐥𝐨𝐫𝐞.
The study highlights the severe limitations of AI in recognizing emotions: the machine confuses mere physical activation with true subjective emotional intensity and is unable to distinguish a genuine emotion from a posed one. Furthermore, to force the use of universal patterns, algorithms discard 70% of our facial expressions as “noise,” thereby erasing the true uniqueness of the individual.
Finally, we address the neuro-legal risks: the uncritical use of AI in forensic profiling threatens fundamental rights (such as the right to remain silent), as the software only evaluates the physiological reaction, ignoring our fundamental neurological capacity for self-control and conscious inhibition (free won’t).
LINK: https://ieeexplore.ieee.org/abstract/document/11579650
