We read with sadness about the tragedy of a pedestrian being killed in a collision with a driverless car being tested by Uber. The reaction was instantaneous – with Uber immediately announcing that the driverless car tests will stop. In our view, that was not a data-driven decision, rather, it was a public perception and PR decision.
Will Uber halt driverless car testing forever? I will make the bold prediction that it will not. After some root cause analysis and after enough time that public reaction dies down, they will continue testing. Of that I am sure.
They have no choice. Their future and only sustainable business model is predicated on operating fleets of driverless cars powered by data and AI as much as they run by solar or petrol.
Statistics suggest that 94% of all crashes are as a result of a driver error, as AI replaces the driver it is reasonable to assume that driver error crashes will reduce. AI is not affected by mood or distracted by phone calls and an analysis of Google’s driverless cars shows a remarkably low accident rate. They obey speed limits at all times, drive at speeds that are far more fuel efficient than humans. They never need a break and remain alert without ever getting tired. Lastly, they reduce costs because they do not need to be paid a wage.
So where is the problem here? It is perception and the fact that even if these vehicles are proven to be safer, the public mistrusts the technology. One fatal crash caused by AI confirms this mistrust and is infinitely more unacceptable to public opinion than the thousands of fatal collisions that are caused by human drivers every day.
Notwithstanding, for companies that see the compelling economic and business drivers for AI applications, there are a number of moral and perception issues that need to be taken into account.
People do not like AI because they fear it will take their jobs or do not like the fact it is already replacing people’s jobs.
People do not like AI because they fear the technology and don’t truly understand it.
People are confused about where liability lies. When something goes wrong, who is legally liable, the AI itself or the company that has implemented the AI?
Uber is at the forefront, dealing with these issues head on right now in a very public forum. The people at Uber understand that they must employ AI technology to re-invent their business to survive – it is deemed to be their only path to profitability. But it’s not a straightforward journey. Setbacks like the terrible tragedy that befell their first collision victim have a significant public sentiment impact. They also need to think about the massive public support they had at liberating many people to find a second income or make a living “working for themselves”. What will happen when they move to an AI/driverless model that has the complete reverse effect of putting all those drivers out of work? Will they be so loved at that time? Will the impact be to push people back to “supporting” taxi drivers?
We don’t have the answers for Uber, but we do feel the context here is important. Many businesses will become uncompetitive if they do not adopt AI and data-driven technology. However, the tech needs to be tempered against public opinion and what consumers want and feel comfortable with. We know that ultimately, there is no stopping progress. But as we implement progress, we can also use data to assess sentiment and readiness amongst customers, consumers and the public at large. The two must go together.