Seven leading artificial intelligence companies will introduce new voluntary safeguards designed to minimize misuse and bias within the emerging technology at an event on Friday at the White House.

President

joe

Biden wants to be joined by Amazon.com executives

incl

Alphabet

incl

meta platforms

incl

Microsoft

corp

and OpenAI, which are among the companies committed to a commitment to transparency and security.

Under the agreement, companies will subject new artificial intelligence systems to internal and external testing before release and ask outside teams to examine their systems for security flaws, discriminatory tendencies, or risks to U.S. rights, health information, or safety.

The companies, including Anthropic and Inflection AI, will also make new commitments to share information to improve risk mitigation with governments, civil society and academics and to report vulnerabilities as they emerge. And leading AI companies will include virtual watermarks in the material they generate, providing a way to distinguish real images and videos from those created by computers.

The package formalizes and expands on some of the moves already underway at major AI companies, which have seen massive public interest in their emerging technology only matched with concerns about the associated societal risks.

Column: Chatbots with artificial intelligence are spreading fast, but the hype around them is spreading faster

Nick Clegg,

the

Meta’s president of global affairs said the voluntary pledges were an important first step in ensuring responsible guardrails for AI are put in place and create a model for other governments to follow.

AI should benefit society as a whole. For that to happen, these powerful new technologies must be built and deployed responsibly, he said in a statement released early Friday.

White House officials say the pledge helps balance the promise of artificial technology with the risks, and is the culmination of months of intense behind-the-scenes lobbying. Many of the executives expected at the White House on Friday attended a meeting with Biden and Vice President Kamala Harris in May where the administration warned the industry that it was responsible for ensuring the safety of its technology.

We need to make sure the companies pressure test their products as they develop them and certainly before they release them to make sure they don’t have any unintended consequences, like being vulnerable to cyber attacks or being used to discriminate against certain people, White House chief of staff Jeff Zients said in an interview. And most importantly, and you’ll see this throughout the work, they can’t check their own homework here.

Voluntary guarantees

Still, the fact that the pledges are voluntary illustrates the limits of what the Biden administration can do to keep the most advanced AI models away from potential abuse.

FTC is investigating ChatGPT about potential consumer harm

The guidelines do not require approval from specific third-party experts to release technologies, and companies are only required to report, rather than eliminate, risks such as potential inappropriate use or bias. The watermarking system has yet to be developed, and it may prove difficult to stamp content in such a way that it cannot be easily removed by malicious actors seeking disinformation on the Internet.

And there are few mechanisms beyond public opinion to enforce commitments to use the technologies for societal priorities such as medicine and climate change.

It’s a moving target, Zients said. So not only do we need to execute and implement these commitments, but we also need to figure out the next set of commitments as the technologies change.

Zients and other government officials also say it will be difficult to keep up with emerging technologies without congressional legislation that both helps the government enforce stricter regulations and allocate funding that allows them to hire experts and regulators.

Employees describe concerns about artificial intelligence as a top priority for the president in recent months. Biden regularly brings up the topic in meetings with economics, national security and health

advisors advisors

and has held talks with cabinet secretaries telling them to prioritize exploring how the technology could intersect with their agencies.

Dating in LA is exhausting, so I asked a chatbot to flirt for me. It almost got weird

In conversations with outside experts, Biden was warned that algorithmic social media such as Metas Facebook and Instagram and TikTok from ByteDance Ltd. have already illustrated some of the risks of artificial intelligence. One outside

advisor advisor

suggested that the president should view the issue as similar to cloning in the 1990s, requiring clear principles and guardrails.

The White House said it had consulted the governments of 20 countries before Friday’s announcement.

I think all sides were willing or excited to get down to business on this as soon as possible because that’s how AI works, you can’t sleep on this technology, Deputy Chief of Staff Bruce Reed said.

However, all of these efforts have lagged behind the pace of AI developments, spurred by intense competition between business rivals and fears that Chinese innovation could overtake Western progress.

Even in Europe, where the EU’s AI law is far ahead of anything passed by the US Congress, leaders have recognized the need for voluntary commitments from companies before binding legislation takes effect. A White House official estimated that it could take at least two years for European regulations to take effect

influencing influence

AI companies.

If you outsource one thing to ChatGPT, job seekers say it should be it

As a result, officials there also asked companies to check themselves. In meetings with tech executives over the past three months, Thierry Breton, the European Union’s Internal Market Commissioner, has called on AI developers to agree to an AI pact to

–

binding guardrails.

–With the help of Bloomberg, writer Jennifer Jacobs contributed to this report.