Due to misuse of AI, people are deprived of benefits and are subject to discrimination: The Guardian

The Guardian's editorial team has conducted an investigation that culminated in a publication accusing government officials of the chaotic and uncontrolled use of artificial intelligence and algorithms to make decisions on numerous critical matters, ranging from benefits allocation to marriage certificate confirmation.

The publication asserts that government officials from at least eight governmental departments and several law enforcement divisions employ AI to make decisions related to social welfare, immigration, and criminal justice. The Guardian contends that such practices lead to discrimination. For example, errors in the algorithm employed by the Department of Work and Pensions (DWP) resulted in dozens of individuals losing their benefits. The facial recognition tool adopted by the London police, under certain settings, makes more mistakes when identifying individuals with darker skin tones.

The Ministry of Internal Affairs has admitted to using AI in passport scanning at airports, assisting in passport application submissions, and detecting fraudulent marriages. Subsequent internal reviews within the ministry revealed that the algorithm employed for detecting sham marriages disproportionately focused on individuals from Albania, Bulgaria, Greece, and Romania.

Bulgarians also suffered the consequences of the DWP algorithm used to identify fraud among benefit claimants. Kate Osamor, a Member of Parliament from the Labour Party, believes that the use of this algorithm may have contributed to dozens of Bulgarians abruptly ceasing to receive benefits as their applications were erroneously identified as potentially fraudulent. Nevertheless, the DWP insists that the algorithm does not consider nationality.

The investigation received commentary from Marion Oswald, a law professor at Northumbria University: "There is a lack of consistency and transparency in how AI is employed in the public sector. Many of these tools affect many individuals in their daily lives, such as those seeking benefits. However, people do not understand why they are used and do not have the means to challenge their application."

Previously, in the United Kingdom, there existed an independent government advisory council that monitored the government's use of AI. However, it was disbanded earlier this year.

The Guardian also reminds readers that issues related to the use of AI by public agencies are not unique to the United Kingdom. In the Netherlands, tax authorities utilized AI tools to detect potential child care benefits fraud. Ultimately, the tax authorities themselves were fined €3.7 million after repeatedly making incorrect decisions and "plunging tens of thousands of families into poverty."