Former Treasury chief adviser warns HMRC close to tracking personal finances with AI


A former senior adviser to Gordon Brown's Treasury has warned that HMRC is about to use artificial intelligence to track the income and expenses of people and businesses without them knowing.

Dr Chris Wales, who was a member of Mr Brown's Council of Economic Advisers for more than six years, has raised the alarm by launching a chilling book co-written by international lawyer Robert Amsterdam on the conduct of the Spanish tax authority. Tax Agency.

He is due to join former Labor Treasury minister Baroness Dawn Primarolo at an event next week to highlight how the Spanish model for tackling tax evasion is about to reach the UK, suggesting the door is opening to a “surveillance state”.

HMRC could follow Spain's example in monitoring personal finances (Reuters)

In a preview of the future, Dr Wales has claimed that confidentiality in personal life – not just finances – will “simply disappear” and asks whether there are adequate safeguards in the UK to prevent HMRC imitating its Spanish counterpart.

He said: “From January 1, all invoices will go through the tax agency in Spain. The inspector can now obtain all your utility bills and will soon find out which clinic and pharmacy you use and what you buy there, what restaurants you eat in, where you buy wine and groceries, what type of car you have, how far you drive and where you park, what flights you take and what hotels you use. Information security? A thing of the past.”

He continued: “I am far from a libertarian, but I see great danger in the direction in which the powers of the tax authorities are going, particularly as the process does not appear to involve our active consent. There is little parliamentary debate about it. In Spain it is simply out of control. In the UK, let's see.”

Highlighting the CONNECT AI program already used by HMRC in the UK, Dr Wales said the UK is now close to following Spain's example.

He said: “HMRC has been using sophisticated information technology for years, including an artificial intelligence system called CONNECT which, as early as 2023, was said to contain more than 55 billion data elements relating to taxpayers.

“Today it will be much greater with these billions of data on taxpayers that AI will be able to classify quickly.”

Dr Wales, who is now senior research adviser at the International Center for Taxation and Development, added that HMRC also refuses to say what algorithms it uses, on the pretext that publishing them will make people “game the system”, a claim he says does not stand up to scrutiny.

“It is understood that the system is used to combat evasion. For the tax authorities, everyone is a potential tax evader. This means that they believe they have a legitimate reason to collect data on all of us,” he said.

Dr Wales and Baroness Primarolo will urge UK MPs to take a stronger stance and apply greater scrutiny than has occurred in Spain.

He noted that the Spanish government is trying to introduce a new law, making it an official secret how data is used, what algorithms are used in selecting taxpayers for research and whether there is any review by an official, in apparent defiance of the EU AI Law, the GDPR and its own Constitution. “This is obviously a cause for deep concern. When the reasons why decisions are made are unknown, legal challenge becomes almost impossible.”

In Spain, authorities are already using the system against British and other expatriates.

An HMRC spokesperson said: “Our collection and data powers are set by Parliament and are subject to strict legal safeguards, oversight and data protection laws. “They exist so that we can collect the right tax to fund vital public services and tackle error and fraud in a way that minimizes intrusion on the honest majority.

“Artificial intelligence underpins some of our processes, but never replaces human decision-making and oversight. We remain committed to the safe use of these technologies, backed by strict ethical, security and data protection standards.”

scroll to top