CONNECT has deployed advanced analytics since the early 2000s, using pattern recognition, predictive modelling and machine learning to identify underpaid tax. The system, developed by BAE Systems Applied Intelligence at an estimated cost of £45–100 million, has helped recover more than £3 billion in unpaid tax. CONNECT reviews multiple data sources, including publicly available social media posts, blogs and other internet content, and can identify patterns and anomalies in seconds, enabling real-time risk profiling and supporting investigations. Outputs are subject to human review and oversight and used alongside investigator work. HMRC's public approach mirrors other departments’ use of online data.
"CONNECT uses (and has always used) advanced analytics such as pattern recognition, predictive modelling, and machine learning, which are all forms of AI. Social media is just one of the many sources CONNECT reviews," she explained.
"CONNECT can identify the patterns and anomalies in the data it reviews in seconds where human investigation would take months," she said. "It not only enables real-time risk profiling; it also supports the work carried out by HMRC staff during the course of investigations."
"The information gleaned and analysed by the CONNECT system is always also looked at by human investigators. As long as there is appropriate human oversight and safeguards, I do not see any problem with the use of AI to identify possible indicators that tax is not being paid at the correct levels."
Collection
[
|
...
]