AI Companies Agree To Voluntary Safeguards, Biden Announces

Seven companies at the forefront of developing rapid advancements in artificial intelligence have agreed to voluntary safeguards for users, the White House announced Friday.

Amazon, Anthropic, Google, Inflection, Meta, Microsoft and OpenAI have all agreed to "voluntary commitments for responsible innovation" that underscore three fundamental principles of "safety, security and trust," President Biden announced after meeting with top executives from the companies.

The emergence of widely available AI tools capable of crafting unique text and images based on user prompts, like OpenAI's ChatGPT chatbot and DALL-E 2 image generator, has sparked an arms race among major tech firms seeking to incorporate similar technology in their own products and advance research in the still-emerging field. Observers say AI has the potential to upend entire industries, but the powerful nature of the technology has also sparked calls from lawmakers — and some of the firms themselves — for more federal regulation to set the rules of the road.

On Friday, Mr. Biden announced several steps that the companies have agreed to take voluntarily.

First, the companies have agreed to "testing the capabilities of their systems, assessing their potential risks, and making the results of these assessments public." They will also safeguard their models against cyberthreats, and manage the risk to national security, Mr. Biden said. Third, the companies "have a duty to earn the people's trust and empower users to make informed decisions, labeling content that has been altered or AI-generated, rooting out bias and discrimination, strengthening privacy protections and shielding children from harm." And finally, the companies "have agreed to find ways for AI to help meet society's greatest challenges, from cancer to climate change," the president said.

The pledges are broad and leave room for interpretation. Some advocates for greater government oversight of AI said the agreements were a good sign, but should still be followed with further regulation.

"These commitments are a step in the right direction, but, as I have said before, we need more than industry commitments. We also need some degree of regulation," said Democratic Sen. Mark Warner of Virginia, the chairman of the Senate Intelligence Committee.

National Security Council spokesperson John Kirby said on MSNBC that the Biden administration is working on an executive order and will pursue legislation to offer guidance on future innovation. 

In October, the White House rolled out what it called a "blueprint" for an AI bill of rights, addressing matters like data privacy.