- By Shiona McCallum
- technology reporter
image source, Getty Images
The White House announced that seven major companies in the artificial intelligence space are working to manage the risks posed by the technology.
This includes testing the security of AI and publishing the test results.
Representatives from Amazon, Anthropic, Google, Inflection, Meta, Microsoft and OpenAI made the announcement alongside US President Joe Biden.
This follows many caveats about the capabilities of this technology.
The pace of each company’s tool development raises concerns about the spread of disinformation, especially in the run-up to the 2024 US presidential election.
Facebook parent company Meta announced Wednesday its own AI tool called Llama 2.
Meta’s global president, Sir Nick Clegg, told the BBC that “the hype is getting a little ahead of the technology.”
As part of the agreement signed on Friday, the companies agreed to:
- Security testing of AI systems is done by internal and external experts before release.
- Make sure people can find your AI by implementing watermarks.
- Regularly and publicly report AI capabilities and limitations.
- Investigate risks such as prejudice, discrimination, and invasion of privacy.
The purpose, according to the White House, is to make it easier for people to tell when online content has been created by AI.
Watermarking AI-generated content was among the topics that EU Commissioner Thierry Breton discussed with OpenAI CEO Sam Altman during a visit to San Francisco in June.
In a tweet that included a video snippet of himself and Altman, Bretton said he was “looking forward to furthering the discussion, especially regarding watermarking.”
In a video clip, Altman said he would “soon” “show” what OpenAI is doing with watermarks.
The voluntary safeguards signed on Friday are a step toward tougher regulation of AI in the United States.
The White House said it would also work with allies to establish an international framework to govern the development and use of AI.
Warnings about the technology include that it could be used to create misinformation, destabilize society, and even pose an existential threat to humanity, but some groundbreaking computer scientists say the apocalyptic warnings are overblown.
