Serious Concerns Raised Over DWP’s Use Of AI To Read Benefit Claimants’ Letters – What It Means For You

The Department for Work and Pensions (DWP) is using artificial intelligence (AI) to process thousands of letters and emails daily through a system called White Mail.

This AI-powered tool is designed to handle vast volumes of correspondence, prioritizing cases deemed most urgent, particularly from vulnerable individuals.

However, concerns are mounting over its lack of transparency and the way it processes highly sensitive personal data without informing claimants.

White Mail

With over 20 million benefit claimants in the UK, the DWP receives around 25,000 letters and emails per day. Traditionally, human staff took weeks to review and sort through this correspondence, causing delays for those in urgent need of assistance.

White Mail, however, can process these communications in a single day, flagging priority cases for human review.

The AI scans and categorizes messages based on urgency and vulnerability, supposedly ensuring that those needing immediate support are seen first.

However, this system also means that some claimants’ letters may be deprioritized, raising questions about fairness and accuracy.

Privacy

One of the biggest issues with White Mail is its handling of sensitive personal data. According to a Freedom of Information (FOI) request, the system processes:

• National Insurance numbers
• Dates of birth
• Addresses and contact details
• Benefit claim details
• Health records
• Bank account information
• Racial and sexual characteristics
• Children’s details, including special needs

Despite dealing with such highly sensitive data, benefit claimants are not informed that their information is being processed by AI. An internal Data Protection Impact Assessment (DPIA) stated that claimants “do not need to know” about their involvement in the initiative.

Furthermore, the DWP has not listed White Mail on the government’s AI transparency register, despite a ministerial mandate requiring public sector algorithms to be logged.

Lack of Consent

Many experts and charities are raising red flags over the ethical implications of White Mail. Meagan Levin, Policy and Public Affairs Manager at Turn2us, said:

“Processing such information without claimants’ knowledge and consent is deeply troubling. The lack of transparency around how the AI makes decisions and handles personal data is a major concern.”

Although the DWP claims that no decisions are made solely by AI, the system still influences which cases are prioritized. It also collects data to “determine insights” and conduct “theme analysis”, though there is little public information about what these insights are used for.

Calls for Oversight

Critics argue that prioritizing some cases means deprioritizing others, which can impact those in need of urgent support. Advocacy groups are calling for greater oversight and transparency, including:

Recommendation	Purpose
Publishing performance data	To ensure fairness and efficiency
Regular audits	To prevent bias and inaccuracies
Appeals process	To allow claimants to challenge AI-driven prioritization
Clear guidelines	To inform claimants how their data is used

Levin emphasized that “transparency and accountability must be at the heart of any AI system”, urging the DWP to provide clear public documentation on how White Mail operates.

Final Thoughts

While AI can improve efficiency in processing benefit claims, the secrecy surrounding White Mail raises serious ethical concerns. Claimants deserve to know how their data is being used, especially when dealing with such sensitive personal information.

Without proper safeguards, the system risks harming rather than helping those who depend on government support.

The DWP has yet to provide an official response regarding these concerns.

FAQs