Why the Business Roundtable just released new guidelines for ‘responsible’ A.I.

A new example of stakeholder capitalism at work is out this morning. The Business Roundtable released guidelines for companies to develop “responsible A.I.,” along with guidance for the U.S. government to develop policies and regulations for A.I. The second piece—advising government on policies and regulations—has long been part of the BRT’s remit. But the first piece—advising companies how to adopt artificial intelligence in a way that maximizes positive social impact and minimizes negative—is new. 

Accenture CEO Julie Sweet led the CEOs in developing the guidelines. I spoke with her on Friday and asked why they are being released now.

“We really believe we are at a tipping point where A.I. at scale is happening. There is true momentum. So we need to act now so that as we deploy at scale, we can act responsibly. You cannot reverse engineer responsible A.I.”

Sweet said the guidelines were developed over a six-month period with input from virtually all the BRT’s 230 member companies. When I asked if any of the companies objected to the guidelines, she responded: “Not to my knowledge.”

“This is a great example of the commitment to living our purpose to serve all stakeholders. It is a defining moment for how companies can collaborate and innovate for the benefit of all.”

The guidelines focus, among other things, on:

• Ensuring A.I. systems are developed by diverse teams, with diverse oversight;
• Employing safeguards to ensure unfair biases aren’t built into A.I. systems, and building A.I. systems that mitigate human bias;
• Designing A.I. systems that are transparent, explainable and interpretable;
• Ensuring A.I. systems are constantly evaluated; that data is collected and used responsibly; that companies have culture and governance systems to support responsible A.I.; and that security safeguards exist from the outset;
• Investing in an A.I.-ready workforce.

That last one, of course, is critical. A.I. has the ability to make work more productive and meaningful, but only if the workforce has the skills needed to use it productively and responsibly. The training challenge is immense—for companies, educational institutions, and governments.

As for policy and regulation, the BRT, not surprisingly, urged a cautious approach, that is industry-specific and use-case-specific. That’s different than the cross-industry approach the EU has taken, as Fortune’s Jeremy Kahn explains smartly here.

Do these BRT recommendations ensure all companies will take a “responsible” approach to developing A.I.? Almost certainly not. The lobbying group has no enforcement tools. But is it good that CEOs are wrestling questions of potential social impact before rolling out A.I. at scale? I think the answer is clearly yes. It wouldn’t have hurt if social media companies had done the same a decade ago.

You can learn more about the BRT guidelines here. Other news below.

Alan Murray
@alansmurray
alan.murray@fortune.com

TOP NEWS

RIP Diem

Meta's big cryptocurrency play, Diem (formerly known as Libra), appears to be toast. With regulators never having warmed to the idea of a Facebook-steered "stablecoin", Mark Zuckerberg's operation is reportedly considering a sale of Diem's assets so as to return capital to investor members. A reminder if it's needed: Meta/Facebook still hasn't managed to build anything significant beyond the social network. Fortune

Bitcoin or bailout?

The IMF won't give El Salvador the loan it's after (it asked for $1.3 billion) until it strips Bitcoin of its legal-tender status—a controversial move on the part of President Nayib Bukele that Fortune's Shawn Tully has skewered here, and while you're at it here's Tully on the Russian central bank's opposition to cryptocurrencies. Fortune

Amazon propaganda

Amazon has quietly dropped a propaganda campaign in which it paid workers to trill on social-media platforms about how wonderful it is to work for Amazon. The "fulfillment center ambassador" scheme was widely mocked over the last few years, not least because there are plenty other people on social media who report it's not so nice working in an Amazon warehouse. Financial Times

Intel fine

Intel has won its 13-year battle to overturn what was in 2009 a record-breaking antitrust fine levied by the European Commission. The case involved "fidelity rebates" that Intel gave PC-makers and retailers to avoid selling products that used chips from key rival AMD. Turns out the Commission should have better explained why these rebates were anti-competitive. Fortune

AROUND THE WATER COOLER

Katie Haun

Venture capitalist Katie Haun, who is leaving Andreessen Horowitz to launch her own fund, is the subject of an exclusive profile by Fortune's Michal Lev-Ram: "'One of the things I learned in the last four years at the firm is I’m actually an entrepreneur,' she says. When a16z’s third dedicated crypto fund was fully deployed last year, it was a natural moment to move on. 'Given that we were at this inflection point, it caused me to do a lot of soul searching,' she says. 'It was time to go out on my own. It’s pretty much that simple.'" Fortune

Google privacy

Google is trying to phase out cookies as a way of giving marketers information about people's online habits. After industry and civil-society pushback, it has now dropped its idea of organizing people into closely-defined "cohorts"—an idea that some warned would allow individuals to be identified from the data—and will instead lump them into very broad categories called Topics, such as "fitness" or "travel". Wall Street Journal

Partygate timeline

If you're having trouble keeping up with all the parties that took place (probably illegally) at the British Prime Minister's residence during hard lockdown, here's a handy timeline of the sordid affair, just in time for the possible downfall of Boris Johnson in the coming days. Fortune

After Omicron

The World Health Organization has warned that the next major variant after Omicron will be even more transmissible, because otherwise how would it displace Omicron? The outstanding question is how severe future variants will be, and on that one the WHO is warning there's "no guarantee" the virus will become weaker as it evolves. Fortune

This edition of CEO Daily was edited by David Meyer.