Talk Is Easy

On June 7, 2018, Google’s CEO Sundar Pichai wrote a blog stating the ethos by which the company would pursue its artificial intelligence pro­gramme. It was titled: “AI at Google: our principles”. On February 4 this year, a line came over it that said Google had updated the principles. There were significant chang­es. Earlier, they had stated that AI would not be used for technologies that could cause overall harm, in weapons and for surveillance. All these went missing in the new principles. One can only infer that Google now does not see its AI programme as a bar for these fronts.

The reason is obviously commercial. Now that the technology has advanced into the real world, there are limitless opportunities and sticking to the original principles would mean being deprived of major markets like the defence sector. But then you have to wonder why those high­falutin aspirations were announced back then. There was an image that the company was trying to present itself to the world—that they were so aware of the dangers of what they were doing, guardrails were being created against it. Noth­ing has changed about that prospect of AI. It will make its way into each and every field that Google had been so prescient about, from weapons to surveillance, and the impact will be immense. Now, however, Google is willing to not just live with it but participate in it.

When it first released the principles, there was no cost attached to it. To affirm it now means a hit on the bottom line. There is also competition. Both human beings and institutions slide down every slippery slope using the argument that if they don’t do it, someone else will—it cannot be stopped, so might as well be at the forefront of it. There is nothing wrong with this position. The jarring note is only in the original statement. It was nothing more than virtue-signalling that fell on its face as soon as it got tested. Then there is the corollary: if the original principles had no sanctity, why should the new updated ones be taken seriously?

AI principlesAI programmeGoogleSundar Pichaitechnology