Can You Trick an AI Lie Detector? [ARTICLE]

New artificial intelligence can find out whether people are lying about who they are and why they are travelling

AiiA Editor

Smarter cities with AI lie detectors

Try to fool an artificial intelligence lie detector, discover innovative fashion made using rare pictures of the Antarctic and find out why 7,000 socks will be hanging in Central Library, as the UK's Manchester Metropolitan University joins the Manchester Science Festival 2018.

The festival, which is in its 12th year, is organised by the Science and Industry Museum – with Manchester Met showcasing some of its most exciting projects and researchers in locations across the city.

RELATED: Smart cities need to be more human, so we're creating Sims-style virtual worlds

Scientists from Manchester Metropolitan will reveal their cutting-edge research, answer any burning questions about how science and technology will shape our futures, and give you the opportunity to take part in live hands-on science experiments.

Are virtual border guards the future?

Interest in and focus on smart cities has skyrocketed in 2018, with a very large number of vendors from across the value chain repositioning and optimizing their IoT portfolios to take advantage of this beckoning opportunity.

Members of the public are invited to attempt to get past iBorderCtrl – a virtual border guard that uses artificial intelligence to work out whether people are lying about who they are and why they are travelling.

Check out this video to the AI border guard in action

iBorderCtrl is a European-funded project that could transform the future of international border security. It is due to be piloted at land borders across the continent later this year – in Hungary, Greece and Latvia.

"The technology can tell if a person is lying or telling the truth through almost imperceptible signs, such as facial micro-gestures"

As part of an advanced trip registration system, travellers are interviewed by an avatar border guard which asks questions about things such as their reasons for travel and how they are travelling to their destination, creating a psychological profile and giving the user a score based on their truthfulness.

RELATED: Global smart cities IoT tech revenues to exceed $60bn by 2026

Behind the scenes, the Automatic Deception Detection system, developed at Manchester Met, can tell if a person is lying or telling the truth through almost imperceptible signs, such as facial micro-gestures.

iBorderCtrl is a ‘human in the loop’ system, so it will give a risk score based on the user’s truthfulness, but human border guards will make any final decisions.

The iBorderCtrl system learns new ways to recognise deception from each person who uses it, so everyone who has a go will be adding crucial – but anonymous — data to the project.

Dr Keeley Crockett, a member of the European consortium developing the iBorderCtrl software at Manchester Metropolitan University, said:

“iBorderCtrl technology uses artificial intelligence to detect micro-gestures and other tell-tale signs of lying – making our borders safer from anyone who intends harm, and speeding up security checks for the majority of travellers.


“It’s great to be part of such a huge and significant project, with partners across Europe, and it is exciting that research and technology developed in the heart of Manchester could revolutionise the future of travel and international security.

“We — the consortium team and myself — are looking forward to meeting the public at the Manchester Science Festival, teaching people about the benefits of this software, and having some fun with it too.”

Check out this podcast on the benefits of implementing RPA into workforce processes