UK Government Urged to Reveal Details on AI Use in Welfare Claims

JJohn September 3, 2023 1:21 PM

The UK government is facing potential legal action unless it becomes more transparent about its use of artificial intelligence (AI) to vet welfare claims, according to the information commissioner. The Department for Work and Pensions' (DWP) refusal to disclose information is raising concerns specifically about the potential impact of the AI system on children and other vulnerable groups.

Concerns rise over AI in welfare claims

Artificial Intelligence (AI) is increasingly playing a role in the UK's welfare system, specifically in vetting claims. The Department for Work and Pensions (DWP) has been deploying machine-learning algorithms over the past two years to identify fraud and errors in universal credit claims. The level of secrecy surrounding this system has alarmed transparency campaign groups and raised serious concerns about the impact on children and other vulnerable groups if benefits are suspended as a result of the AI's decision making. The information commissioner has warned the DWP of potential legal repercussions unless it improves its transparency about the use of AI.

DWP criticized for lack of transparency

The DWP has faced criticism for repeatedly refusing to disclose details about its AI system. This includes information about the data used to train the AI, the companies involved in its development, and an analysis of any potential biases that could unfairly impact certain groups. The lack of transparency has been described as 'seriously concerning', especially as the system is known to have shown some evidence of bias in trials. The DWP's refusal to answer questions about how historical fraud and error data is used in training the AI model is particularly contentious.

Potential legal action for DWP

The information commissioner, John Edwards, has issued a stern warning to the DWP, stating it could be in contempt of court unless it changes its approach towards transparency. Edwards has given the DWP a 35-day deadline to clarify the terms on which it could release more information about the AI system. The DWP's actions, including its refusal to comply with freedom of information requests and its changing reasons for not releasing information, have been criticized as 'unfortunate' and 'incorrect'.

As the UK continues to digitize its welfare system, relying more heavily on machine learning and AI, concerns are growing about the risks for benefit claimants. Without greater transparency, there is the potential to create serious problems for some benefit claimants, particularly if the system is biased or erroneous. Campaign groups and international experts have underscored the importance of public transparency, especially when powerful technologies with potentially significant impacts are in use.

The DWP's overall handling of freedom of information requests has come under fire. The information commissioner has previously warned the DWP about its consistently poor performance in this area, and has taken issue with its tendency to use standard templates when refusing requests. The department has also been accused of withholding more information in recent times. As a result, the DWP has been ordered to outline how it might be able to respond to a specific request for information by late September.

More articles

Also read

Here are some interesting articles on other sites from our network.