By akademiotoelektronik, 04/12/2022

Artificial intelligence: Google launches a new competition endowed with 25 million dollars

Google wants to help people do good things with artificial intelligence (AI). The California-based company on Monday announced a new competition for nonprofits, universities and others working on AI projects that will benefit society. Titled the AI ​​Global Impact Challenge, this competition has $25 million in grants. It is part of a new initiative from Google called “AI for Social Good”. This program is intended to help address urgent issues related to crisis relief, environmental conservation, or sex trafficking.

But the project also comes at a time when Google is under increased scrutiny over how its own artificial intelligence might be used, including in controversial military work or efforts to build an engine. censorship-compliant research in China. During a press conference, Jeff Dean, head of artificial intelligence at Google, did not directly address these controversies, but he did refer to the company's ethical principles for artificial intelligence, which describe how it intends to use this technology or not. “We are all grappling with questions about how artificial intelligence should be used,” he said. “AI really has the potential to improve people's lives. »

Competition winners will also have access to Google's technical resources and an in-house expert will be appointed to help them develop their projects. The company has opened the application process and will announce the winners next spring at the annual Google I/O developer conference. “The core of the program is to encourage people to take advantage of our technology. Google can't work on everything,” explained Yossi Matias, vice president of engineering. “There are a lot of issues that we may not even be aware of. »

Google employees take a stand

At the same time, Google has faced backlash against some of its own artificial intelligence projects. The company's cloud division, under the leadership of Diane Greene, has embarked on securing lucrative military contracts. But employees have challenged Google's decision to participate in Project Maven, a US Department of Defense program to develop better AI for the US military. More than 4,000 employees are said to have signed a petition addressed to Sundar Pichai, CEO of Google, asking him to cancel the project. In June, Google said it would not renew Maven's contract and would not seek similar contracts. A week later, Sundar Pichai published a text detailing the ethical guidelines for the development of AI. He assured that Google would not create technology that would be used to make weapons, but would continue to work with the military.

The company has also been criticized for Duplex, an artificial intelligence that can make natural language phone calls to reserve a table at a restaurant or book a hair appointment through the Google Assistant . Duplex's hyper-realistic voice, which even goes so far as to play on pause times between responses and uses verbal tics like "um" and "uh", has raised concerns that she might mislead her interlocutors. Google was quick to indicate that the system would specify that it was a robocall at the start of each conversation.

Still, the AI ​​for Social Good program is not a response to recent controversies, assures Yossi Matias that its gestation began a long time ago.