Can we trust AI if we don't know how it works?

[Visualmodo Team]

We’re at an unprecedented point in human history where artificially intelligent machines could soon be making decisions that affect many aspects of our lives. But what if we don’t know how they reached their decisions? Would it matter?

Imagine being refused health insurance – but when you ask why, the company simply blames its risk assessment algorithm.

Or if you apply for a mortgage and are refused, but the bank can’t tell you exactly why.

Or more seriously, if the police start arresting people on suspicion of planning a crime solely based on a predictive model informed by a data-crunching supercomputer.

These are some of the scenarios the tech industry is worrying about as artificial intelligence (AI) marches inexorably onwards, infiltrating more and more aspects of our lives.

AI is being experimented with in most sectors, including medical research and diagnosis, driverless vehicles, national surveillance, military targeting of opponents, and criminal sentencing.

A recent report by consultancy PwC forecasts that AI could boost the global economy by $15.7tn (£11.7tn) by 2030.

But at what cost? These software algorithms are becoming so complex even their creators don’t always understand how they came up with the answers they did.

[Author]

Posts