# L'algoritmo è uguale per tutti? | Ernesto Belisario | TEDxLink Campus University Source: https://www.youtube.com/watch?v=GeyVWj2Ho7c Recap page: https://rapidrecap.app/video/GeyVWj2Ho7c Generated: 2025-12-03T20:05:04.283+00:00 --- ## Quick Overview The algorithm is not yet equal for everyone in the justice system, as demonstrated by studies showing AI bias against minorities, but the speaker argues that carefully designed AI can help make justice systems less arbitrary, less unjust, and more truly equal by augmenting human decision-making. **Key Points:** - The legal maxim "The law is equal for everyone" is currently challenged by the use of AI in judicial settings, as illustrated by a Florida case where a Black woman received a high-risk score (8) while a white man received a low-risk score (3) for similar crimes. - A study by researchers at Harvard found that AI risk assessment tools, like the one used in Florida, systematically show bias, disproportionately flagging minorities as high-risk for recidivism. - The French legal system passed a law in 2019 that explicitly forbids the use of generative AI (like ChatGPT) in judicial activities, including interpretation of law, fact evaluation, and adoption of rulings. - The speaker points to the existence of a database tracking "AI Hallucination Cases," which documented 597 instances of fabricated legal citations or arguments generated by AI in court filings. - The speaker suggests that the solution is not to ban AI entirely, but to use it as a tool to support human judgment, making the system less arbitrary and more equitable. - The bias in AI systems stems from the data they are trained on, often reflecting existing societal biases, such as those prevalent in North American and Western cultural profiles. ![Screenshot at 03:11: The speaker displays a slide comparing two individuals assessed by a risk assessment algorithm: a Black woman scored 'HIGH RISK 8' and a white man scored 'LOW RISK 3', illustrating the inherent racial bias in the AI's output.](https://ss.rapidrecap.app/screens/GeyVWj2Ho7c/00-03-11.png) **Context:** This TEDx talk, titled "L'algoritmo è uguale per tutti?" (Is the algorithm equal for everyone?), is delivered by Ernesto Belisario, an attorney from the Ilexia Legal Studio. The presentation explores the ethical and practical implications of integrating Artificial Intelligence, particularly large language models like ChatGPT, into the justice system, contrasting the ideal of legal equality with the reality of algorithmic bias demonstrated in risk assessment tools. ## Detailed Analysis Ernesto Belisario opens by questioning the common legal phrase, "The law is equal for everyone," in the context of AI usage in courts. He immediately presents a stark example from Florida in 2014 where an AI risk assessment tool assigned a Black woman a significantly higher risk score for recidivism (8) than a white man (3) for similar offenses, demonstrating systemic bias rooted in the training data. Belisario references a study by researchers, likely from Harvard (as shown later in a slide referencing a journal article), which tracked these disparities. He also cites the 597 documented cases of "AI Hallucination"—where generative AI fabricated legal citations—as evidence that these tools are not infallible. Furthermore, he notes that France explicitly forbade the use of generative AI in judicial activities in 2019 via a specific law (Article 15). Belisario contrasts the common law system (US), which relies heavily on precedent (case law), with the civil law system (Europe/Italy), suggesting that the former might be more susceptible to biased algorithmic recommendations due to its reliance on historical data. He concludes that while the current iteration of AI reflects and reinforces existing societal biases (often skewed toward Western, affluent cultural profiles), the technology itself is not inherently flawed. The true potential lies in using AI to augment human decision-making, helping judges to make the system less arbitrary, less unjust, and ultimately, more truly equal. ### Introduction and The Core Question - Introducing the central theme of algorithmic fairness in the justice system - Highlighting the contrast between the ideal of legal equality and the reality of AI bias - Mentioning the Florida case demonstrating racial disparity in risk scores (Black woman: 8, White man: 3) ### Evidence of AI Failure - Citing the 'AI Hallucination Cases' database tracking 597 instances of fabricated legal citations - Discussing the French legal system's 2019 law (Art. 15) explicitly reserving final decision-making power to magistrates, prohibiting AI from interpreting law or evaluating facts ### The Nature of Bias - Explaining that AI bias stems from training data reflecting cultural profiles, often favoring Western/North American data - Contrasting Civil Law (code-based) versus Common Law (precedent-based) systems and their respective vulnerabilities to historical data biases ### Conclusion and Path Forward - Posing the question of whether the algorithm can be truly equal - Concluding that AI should serve as an augmentation tool for human judges, not a replacement, to make justice less casual, less unjust, and more equitable ![Screenshot at 00:04: The opening title card displaying the event branding 'LEX MACHINA'.](https://ss.rapidrecap.app/screens/GeyVWj2Ho7c/00-00-04.png) ![Screenshot at 00:39: Slide introducing the speaker, Ernesto Belisario, and the talk's central question: 'L'algoritmo è uguale per tutti?' \(Is the algorithm equal for everyone?\).](https://ss.rapidrecap.app/screens/GeyVWj2Ho7c/00-00-39.png) ![Screenshot at 03:12: Visual comparison slide showing the recidivism risk scores assigned by an algorithm to a Black woman \(HIGH RISK 8\) versus a white man \(LOW RISK 3\).](https://ss.rapidrecap.app/screens/GeyVWj2Ho7c/00-03-12.png) ![Screenshot at 09:57: Slide displaying Article 15 of Italian law concerning the use of AI in judicial activity, stating that final decisions must be reserved for the magistrate.](https://ss.rapidrecap.app/screens/GeyVWj2Ho7c/00-09-57.png) ![Screenshot at 11:09: Illustration contrasting Civil Law \(relying on a single 'Code Civil'\) versus Common Law \(relying on stacks of 'Case Law' and 'Precedents'\).](https://ss.rapidrecap.app/screens/GeyVWj2Ho7c/00-11-09.png)