Alejandro Mayorkas, the US Secretary of Homeland Security, announced to journalists on Tuesday that the department is working on testing artificial intelligence (AI) techniques to train police officers conducting interviews with asylum seekers to the United States.
Mayorkas explained that the Department of Homeland Security is training machines in this pilot project to behave like refugees so that police officers can practice conducting interviews with them.
He added that due to the trauma they have experienced, asylum seekers often refrain from talking about that trauma; therefore they are teaching the machine to be cautious as well, and to adopt “other qualities” of asylum seekers.
Mayorkas’s statements, made on the sidelines of a security-focused conference in San Francisco, highlight the artificial intelligence initiatives previously disclosed by the Department of Homeland Security earlier this year.
According to the department, it plans to develop an interactive application to complete its training of immigration officials, relying on what is known as generative artificial intelligence that provides new content based on previous data.
The department pointed out that the US Citizenship and Immigration Services, specifically a department under the Department of Homeland Security, will work on building an artificial intelligence application to design training materials for the needs of police officers and prepare them to make more accurate decisions.
The department added that artificial intelligence techniques will not make decisions themselves regarding immigration.
According to Mayorkas, these techniques will have the ability to determine the specific conditions of each country and other information that will help facilitate the process.
With this pilot project and several experiments in this field, the government aims to reduce costs and improve performance through artificial intelligence, especially after the significant launch of the Chat GPT platform in 2022.
