The White House announced Thursday a series of measures aimed at better regulating the use of artificial intelligence (AI) in public administrations and “identifying and managing potential risks”, while ensuring “transparency” of uses. . She has developed a new set of standards aimed at “protecting” the security of Americans’ data and ensuring that they “know how their government is integrating AI,” Vice-President Kamala Harris said during of an online press conference.
Furthermore, administrations will have to put in place “internal supervision, with the designation of an AI manager with the experience, expertise and authority” to ensure the proper use of the tools. employees, Ms. Harris added. These measures were formulated “after consultation” with officials from the private and public sectors, computer researchers and even human rights specialists, insisted the American vice-president.
Assess the dangers
Among the elements taken into account are possible discriminatory biases, often denounced by associations for the defense of minorities who point out that AI is very often developed by white men from a higher social category. President Joe Biden signed a decree in October intended to better take into account the risks induced by the development of AI. The first measures aimed to assess the dangers, particularly in terms of national security, economic security and public health, and – already – to limit potential discriminatory bias.
Research into the impact of AI on the labor market had also been launched. Apart from uses linked to administrations, the American presidency, however, only has limited room for maneuver: real legislation on the subject would require a vote in Congress, which appears difficult to obtain in this election year. The regulation of AI is on the agenda in several countries and international bodies. In its first resolution on the subject, the UN General Assembly called on March 22 for the establishment of international “standards” to guarantee “safe” tools that respect human rights.
On March 13, the European Parliament approved legislation aimed at promoting innovations in the field while limiting the risks of abuses, legislation which goes further than those implemented so far by China or the United States.