Google publicly announces its self-driving car project

Before the announcement

For most of its life Google was understood as a search company, and public discussion of the company was mostly about the web. The self-driving car research pointed somewhere else. The New York Times, reporting on the project, called it proof that Google's ambitions reached beyond the search engine business ^[2]. The work had not been announced beforehand: the Times said Google had been developing the vehicles in secret but in plain view ^[2], meaning that cars were being driven and tested while the company said nothing about them publicly.

What Google said

In October 2010 Google used its corporate blog to announce that it had developed technology for cars that can drive themselves ^[1]. The post was written by Sebastian Thrun, and it gave the project a three-part motivation: to help prevent traffic accidents, to free up people's time, and to reduce carbon emissions by fundamentally changing car use ^[1]. Safety, time and emissions — that framing presented self-driving as an answer to problems created by ordinary driving, rather than as a technology demonstration for its own sake.

Thrun was not a conventional car executive. The Times described the project as his brainchild and identified him as a 43-year-old director of the Stanford Artificial Intelligence Laboratory, a Google engineer and a co-inventor of Street View ^[2]. The people and the company behind the announcement therefore came from computer science and mapping, not from the automobile industry.

How the cars drove

The Times reported that the vehicles used artificial-intelligence software that senses anything near the car and mimics a human driver's decisions ^[2]. That sentence contains two separate jobs. Sensing means the car has to build a picture of what is around it — where the road is, where other vehicles and people are. Deciding means the software has to choose what a person would choose: when to slow down, when to change direction, when to stop. The first is a problem of perception, the second is closer to writing driving judgement into a program.

Engineers on the project argued that robot drivers react faster than humans, have 360-degree perception and do not get distracted, sleepy or intoxicated ^[2]. Each of these claims is about the weaknesses of human driving rather than about machine intelligence as such. A driver that never looks away, never tires and is never impaired addresses precisely the conditions under which many crashes happen, which is why the safety argument carried weight.

The caveat

Google's own post was careful. It said the technology was still very much in the experimental stage, and presented the work as a glimpse of what transportation might look like thanks to advanced computer science ^[1]. The announcement was therefore a statement of direction rather than the launch of something a customer could obtain.

Why it mattered

The announcement carried weight for two reasons. The first was the source. A company whose business was organizing information was now putting engineering effort into how people and goods move, which is what the Times meant when it read the program as evidence of ambitions beyond search ^[2]. The second was the framing. By naming traffic accidents, wasted time and carbon emissions as the goals, Google placed a computer-science research program inside a public argument about road safety and transport. The project was described as experimental, and the cars were already sensing their surroundings and making decisions on the road — a combination that made self-driving legible to a general audience as something under development rather than as science fiction ^[1][2].