How AI Threatens Your Privacy

Imagine a future where you are unsuspectingly queuing up at a bank. Suddenly the guards show up and remove you from the line, claiming that you are just wasting your time, that you won’t get a loan renewal. Confused about how they know about your intentions, you ask them why and their answer is shocking: apparently an artificial intelligence has been profiling every client that entered the building, checking their background, leniency and financial status. It spotted your low credit score and pre-emptively denied your attempt for a loan.

Does the AI care that your poor credit score is the result of a mortgage you needed to support your parents but otherwise always pay your bills on time? No. Does the bank care? No, they are happy with the system even if it might be faulty since it frees up time and money by vetting people before they reach the clerks. No one cares, just you.

Considering the paths that both AI implementation and facial recognition technology are heading down, this dystopian horror story one day might become reality.

LineSight: a border guard’s best friend

IT company Unisys is currently hard at work on a system that could greatly improve the tasks of border guards and airport security, and yet may possibly inconvenience the life of travelers. The purpose of the application — dubbed LineSight — is to process people’s information and flag suspicious individuals. It would be able launch a deep background search by checking the person’s travel history, cargo manifest, criminal background, traveling behavior (seat selection, check-in intervals etc.) and even see if they are accompanying a person that is also flagged. For this reason, LineSight pulls data from public files, travel agencies and even from Interpol.

The system is capable of autonomously assigning different clearance levels to passengers, prompting the security team to scrutinize someone with an anomaly in their records. Despite not being a full-fledged AI, LineSight uses machine learning to refine its analytical algorithm. The companies using the system can also customize its sensitivity and decide which sources it should use for the investigation; it won’t be difficult to implement a feature that enables the scanning of social media profiles too.

Privacy concerns on the horizon

It’s important to note that the system is far from being an actual artificial intelligence, and admittedly it won’t have absolute control over the vetting process. Still, it’s not hard to see the parallel between the situation presented in the opening of this page and what might become of LineSight in the future. People are already concerned about AIs: according to a survey conducted in China, 80% of their citizens feel threatened by them. The majority are afraid of losing their jobs, while others see a threat to their privacy in the snooping algorithm.

Now imagine these analytical features married with facial recognition. Advanced biometric scanning is already worrisome since it can happen without the target’s consent. A camera at the airport entrance captures your face, identifies you and greenlights the background checking process by LineSight. It’s convenient for the company, and increases the efficiency of the security team. And even if the system makes a mistake, they just chalk it up as just ‘another lesson for the learning algorithm’.

Limiting AI usage to protect privacy