(Credit: Shutterstock/Supatman.)
Several of the world's largest tech firms—including Microsoft and Nvidia—have agreed to join the US AI Safety Institute Consortium, the US Department of Commerce announced Thursday.
The new group was formed as part of President Biden's October executive order calling for AI regulation and has over 200 member companies and organizations, from US universities to multinational tech and financial firms.
The large collective will take on the task of developing national guidelines for AI watermarks, safety, security, and red teaming, to name a few. Red teaming is when an entity poses as its adversaries or possible threat actors by hiring "ethical" hackers, for example, to expose potential vulnerabilities or issues with a technology.
In a statement, US Secretary of Commerce Gina Raimondo said AI comes with risks, but has "immense potential."
"We will ensure America is at the front of the pack," Raimondo said. "Together we can confront these challenges to develop the measurements and standards we need to maintain America's competitive edge and develop AI responsibly."
Other consortium members include Adobe, Amazon, Apple, GitHub, Hewlett Packard, IBM, Linux, Meta, and Qualcomm. AI firms Anthropic and OpenAI have also agreed to join, according to the Department of Commerce's list.
MIT, Cornell, Princeton, Duke, Johns Hopkins, NYU, Georgia Tech, and USC are among the many US universities that have signed on.
"To keep pace with AI, we have to move fast and make sure everyone—from the government to the private sector to academia—is rowing in the same direction," said White House Deputy Chief of Staff Bruce Reed in a statement. "Thanks to Biden's landmark executive order, the AI Safety Consortium provides a critical forum for all of us to work together to seize the promise and manage the risks posed by AI."
Some tech firms are already independently taking steps toward labeling AI-generated content. Last year, Google and OpenAI said they would find ways to mark AI-generated content.
This week, Meta and OpenAI shared their respective efforts to label images that have been AI-generated. While imperfect, Meta's AI labels and OpenAI's metadata additions are first steps toward increasing transparency around AI-generated content as deepfakes and AI art continues to spread online.