A U.N. working party will adopt guidelines as soon as November for preventing cyberattacks on network systems used in self-driving vehicles, a source involved in the negotiations said.
The guidelines will call on automakers to introduce specific measures for alerting drivers when attacks are detected and preventing the loss of control over the vehicles, according to the source.
Discussions on the topic at the World Forum for Harmonization of Vehicle Regulations in Geneva has been led by Japan and Germany, while the United States is planning to establish its own safety regulations separate from the U.N. measures.
Autonomous driving, which frees drivers from controlling a vehicle, involves artificial intelligence technology and requires use of a network of electronic devices that collect and exchange data known as the internet of things.
But such networks have proven vulnerable to hackers trying to take over self-driving vehicles.
U.S. and European media have reported researchers being able to hack a Tesla Motors electric vehicle and remotely apply its parking brakes while it was moving, as well as switch its instrument panel displays on and off.
The guidelines will call for steps to prevent the breach of automated driving systems through the internet and prevent the loss of control over vehicles even if a breach occurs, the source said.
In the United States, the debate over safety standards for self-driving vehicles has been overshadowed by a fatal accident in May involving a Tesla vehicle equipped with an autopilot system for highway driving.
The driver of the partially automated Tesla vehicle died after a collision with a semi-trailer. Tesla said the autopilot sensors failed to distinguish the white semi-trailer against a bright sky and neither the system nor the driver applied the brakes.