Newsroom

One of the problems with the driverless car is that it is highly skilled at self-preservation: all its intelligence is focused on stopping it from crashing and harming its passengers.

Nearby bikes and pedestrians? Collateral damage.

Roads are social systems and rely on ethical decisions by their users to function with safety.

While bike riders have enough experience with drivers to understand the weakness of that idea, autonomous cars – by removing the driver from the seat – promise to reduce driver error and so reduce risk.

A question remained, however: how would the algorithms prioritise survival in a crash scenario? Would bike riders be just another piece of frangible road furniture helping to absorb the energy of a run-away truck?

Such was the concern that the risk was being distributed to favour vehicles and people in cars over those people outside of cars, that the EU Commission stipulated 20 ethics considerations to more fairly balance such risks.

Now, for the first time, there is autonomous driving software that claims to meet these requirements by assessing the risks to bike riders and pedestrians in plotting its vehicles trajectory.

The code, developed at Technical University of Munich, has been open-sourced and made available to the industry.

Maximilian Geisslinger, a scientist at the TUM  and Chair of Automotive Technology, explains the approach: "Until now, autonomous vehicles were always faced with an either/or choice when encountering an ethical decision.

"But street traffic can't necessarily be divided into clear-cut, black and white situations; much more, the countless gray shades in between have to be considered as well.

"Our algorithm weighs various risks and makes an ethical choice from among thousands of possible behaviours – and does so in a matter of only a fraction of a second.”

Approximately 2,000 scenarios involving critical situations were tested, distributed across various types of streets and regions such as Europe, the USA and China.

The implementation incorporates a combination of five essential ethical principles: minimization of the overall risk, priority for the worst-off, equal treatment of people, responsibility, and maximum acceptable risk.

The developed ethical principles do not only focus on potentially hazardous situations but apply to all kinds of driving situations. They ensure transparent decision-making and, by considering responsibility, potentially account for long-term effects.

The settings are modifiable. The researchers say different countries, cultures, or even individuals may demand different ethical emphases.

That could be an opportunity for an Elon-gated ego to flip the algorithm and market the most ethically selfish car. It might sell lots.

The research work published in the journal Nature Machine Intelligence.