Google’s Sundar Pichai has pledged to collaborate with European politicians on an “A.I. Pact”—a voluntary set of guidelines or standards while formal A.I. regulations are being developed.
“There is no time to lose in the A.I. race to build a safe online environment,” said E.U. internal market commissioner Thierry Breton after Pichai’s meeting.
His office briefed following the meeting that the E.U. wants to be “proactive” and work on an A.I. treaty before upcoming E.U. A.I. legislation. The memo stated that the group intends to create an A.I. Pact “involving all major European and non-European A.I. actors voluntarily” before the pan-EU A.I. Act’s deadline.
However, only Google has publicly supported the idea. We asked Google and the European Commission about the initiative.
Breton stated: European technology must follow our data protection, online safety, and artificial intelligence laws. Europe doesn’t choose. I’m glad Sundar Pichai recognizes this and will follow all E.U. legislation.
GDPR applies. Implementing the DSA and DMA. The European Parliament and Council should adopt the A.I. Act framework before the end of the year as negotiations near completion. Sundar and I agreed that we could not wait till A.I. regulation becomes applicable and to engage with all A.I. developers to form an A.I. Pact voluntarily before the legal date. I especially applaud Sundar’s pledge to battle disinformation before European elections.
As with any self-regulatory agreement, the “A.I. pact” would have no legal power, so developers could not be forced to join up or punished for not meeting their voluntary pledges. Still, it may be a step toward the worldwide rule-making cooperation technologists have called for in recent weeks and months.
The E.U. has previously gotten tech giants to sign voluntary agreements (aka Codes) to improve their responses to online hate speech and harmful disinformation. While the two Codes haven’t solved online speech moderation difficulties, they’ve given the E.U. a tool to evaluate platforms’ promises and publicly shame those that don’t.
The E.U. led the world in digital rule-making and suggested a risk-based framework for A.I. apps two years ago. However, even the bloc’s best efforts are still lagging improvements in the field, which have felt especially searing this year after OpenAI’s generative A.I. chatbot, ChatGPT, went viral.
The European Parliament and Council are still debating the April 2021 draft E.U. A.I. Act, which includes multiple generative A.I. modifications. As a result, the E.U.’s A.I. rulebook’s final form will depend on the E.U. co-legislators compromise.
Even if the law is passed before the end of the year—the most hopeful scenario—it will likely take at least a year to apply to A.I. developers. Thus, E.U. commissioners strongly favor interim measures.
EVP Margrethe Vestager, who leads the bloc’s digital policy, stated this week that the E.U. and U.S. would work on baseline criteria before the law takes effect (via Reuters).
After meeting with Google, she tweeted: “We need the A.I. Act as soon as possible, But A.I. technology evolves at extreme speed. We need voluntary agreement on universal A.I. norms now.”
Despite these unexpected demonstrations of high-level haste, the E.U.’s data protection legislation, the GDPR, may apply—and has already been applied against ChatGPT, Replika, and Clearview A.I. For instance, a regulation intervention on ChatGPT in Italy at the end of March briefing led to a service suspension. In addition, OpenAI created new disclosures and controls for users to comply with privacy requirements.
As Breton notes, the incoming DSA and DMA may create additional hard requirements for A.I. app makers, depending on their service, in the coming months and years as those rules apply to digital services, platforms, and the most market-shaping tech giants (in the case of the DMA).
The E.U. believes A.I. needs risk-based guidelines. It appears to want to capitalize on the “Brussels effect” its digital lawmaking can have by establishing an interim A.I. accord.
A Senate committee recently heard from OpenAI CEO Sam Altman on regulating A.I. Google may rush to collaborate with the E.U. on voluntary norms to play the other side. A.I. regulation wars begin!