For years Google published its scientific research, which greatly helped its competitors. But now it has gone into defensive mode.
In February, Jeff Dean, the longtime head of Google's AI division, announced a surprising policy change to his staff:
They should stop sharing their work with the outside world.
For years Jeff Dean ran his department like a university, encouraging researchers to publish their academic work. So they made available almost 500 studies as of 2019, according to the Google Research website.
But the release of OpenAI's groundbreaking ChatGPT changed things. The San Francisco startup keeps up with Google by reading his team's scientific papers, Dean said at the quarterly meeting for the company's research division. Indeed, transformers for example are a fundamental part of the latest AI technology and the T in ChatGPT (Chat Generative Pre-trained Transformer), came from a Google study.
Things had to change. That's why Google will henceforth exploit its own AI discoveries and share its studies only after turning lab work into products, Dean said.
The policy change is part of a larger change within Google. which has long been considered the leader in artificial intelligence. The company has gone into defensive mode — first to fend off a slew of nimble AI competitors, but also to protect its search engine, its stock price and, potentially, its future, which executives say is intertwined with AI. .
In op-eds, podcasts and TV appearances, Google CEO Sundar Pichai says we should be careful with artificial intelligence.
"On a societal scale, it can do a lot of damage," he warned on "60 Minutes" in April, describing how the technology could accelerate the creation of fake images and videos. However, in recent months, Google (as does Microsoft) doesn't seem to be paying much attention, aiming to launch new products fairly quickly.