AI and predictive policing: an ethical dilemma in the digital age

The landscape of ethics in Artificial Intelligence (AI) is marked by lively debate. One of the most controversial aspects concerns the use of AI in sectors that could have a significant impact on human rights, as in the case of Predictive Policing.

This term refers to the process of collecting and analyzing data about past crimes to identify and statistically predict areas or individuals most likely to have criminal activity. In other words, it is the use of algorithms to evaluate the probability of committing crimes in a specific geographical area or the potential danger of individuals, based on the historical data available.

However, authoritative studies have shown that such an approach could lead to discrimination, disproportionately affecting members of ethnic minorities or already marginalized communities. In the United States, for example, the absence of specific regulation has allowed the use of AI for various purposes, including border monitoring and criminal investigations. Only a few cities, such as Portland and San Francisco, have enacted bans on the use of facial recognition in public spaces.

Now, it appears that the US government is trying to increase control over AI. This will is demonstrated by recent meetings in the White House with representatives of Big Tech, whose goal is "to innovate responsibly and protect the rights and safety of people". During a congressional hearing, Sam Altman, CEO of OpenAI, stressed the need for high security standards for the most powerful AI models and close international cooperation in this field.

Geoffrey Hinton, a leading figure in the field of AI, has expressed similar concerns. After leaving Google, he warned that AI could become powerful enough to destabilize the job market and challenge the very concept of "truth". These concerns have been reinforced by researchers such as Timnit Gebru, who has called for more attention to be paid to the ethical implications of these technologies.

Growing concern about these issues has prompted workers in various sectors to demand regulations and protections. On the other hand, the European Union is already actively working to develop specific regulation for AI.

After ENISA's report on AI security standards, the next step will be the Artificial Intelligence Act (AIA). The draft of this act, approved last May 11, will face a first parliamentary examination in the session of June 12-15.

The aim of the new EU rules is to ensure that AI systems are "safe, transparent, traceable, non-discriminatory and environmentally friendly". To achieve this, the EU is adopting a risk-based approach and plans to ban practices considered too invasive, such as "biometric surveillance" and "predictive policing" schemes.

If the AIA text is approved in its current formulation, the use of Predictive Policing will be effectively excluded in Europe.

In conclusion, beyond the technical challenges, the big bet in terms of AI seems to have to be played on the ethical, political and protection of fundamental human rights fronts.