Data Science Investment Counter —Funding raised by UK data science companies in 2018.
£ 5.640 Million

AI tool used to catch people who lie to the police

Scientists at Cardiff University have developed a new computer programme that can spot if someone has lied to police about being robbed.

The new software is called VeriPol, and it analyses the phrasing of a victim’s statement in order to identify particular signs of fake reports.

The technology has been tested by Spanish police with success rates higher than 80 per cent. Through VeriPol, 64 false reports have been identified in just one week.

VeriPol works through a combination of automatic text analysis and artificial intelligence to recognise when somebody has been lying or overstating to the police.

The reasons behind VeriPol’s development are to be found in the thousands of false reports that are submitted to the police each year. Many wrongdoers lie in police reports hoping to receive inflated insurance pay-outs or submitting claims for crimes that never happened to begin with.

But according to the researchers, there are certain revealing signs contained in written statements that the new software can quickly identify.

For example, it is believed false statements are more likely to be shorter than genuine ones and focus on the details of the stolen property rather than the incident itself. They are also usually lacking in certain details about the perpetrator and the attack itself.

The team behind the technology said it could help police forces save valuable time and effort in investigating false crimes and will also help deter people from filing false reports in the first place.

Dr Jose Camacho-Collados, from Cardiff University’s School of Computer Science and Informatics, who helped develop the technology, said: “Our study has given us a fascinating insight into how people lie to the police, and a tool that can be used to deter people from doing so in the future.

“Police officers across Spain are now using VeriPol and integrating it into their working practices. Ultimately we hope that by showing that automatic detection is possible it will deter people from lying to the police in the first instance.”


Co-working space and blog dedicated to all things data science.

Subscribe to our newsletter