Meta's executive director, Mark Zuckerberg, rejected the growing calls to suspend or slow down the development of artificial intelligence (IA), arguing that market mechanisms and risk of lawsuits already function as major brakes for possible abuses.
© Getty Images
16/09/2026
by
Lusa
Na text published on social network X, Zuckerberg said that “people will not want to use agents who are not aligned with their interests, or who do not do what they are asked to do,” so that programmers have “a strong natural incentive” to make models more faithful to users’ needs.
The concept of “alignment” refers precisely to the effort for systems to behave according to human will.
The founder of Meta stressed that any laboratory that does not focus on this alignment “will stay behind” and recalled that the company postponed for months the launch of Muse AI system, with custom agents, to strengthen security mechanisms. “We did it as part of our daily work because it was clearly the right thing for people and for us,” he wrote.
The discussion on AI regulation has intensified in recent weeks, following safety incidents and signs that models begin to evolve without human intervention. Companies such as OpenAI, Anthropic and Google study the creation of a common body to define standards, while figures like Dario Amodei, executive director of Anthropic, advocate a slowdown in the pace of development.
Zuckerberg counterpoints that the threat of legal proceedings already obliges laboratories to act responsibly and that the best practice of industry is independent evaluation of systems.
He added that Meta resorts to external experts and called for the creation of a “widest and more diversified system of evaluators”, in an indirect criticism of Anthropic, accused of favoring peer reviewers close to the company.
The Goal, which this year invested up to $145 billion (EUR 125 billion) in AI infrastructure, has resisted any regulation that might delay its models. Zuckerberg had already argued in August that “no rule should delay the departure of American models, not even for a month.”.
His position was welcomed by several Silicon Valley personalities close to the White House, who contest the warnings about the risks of the AI and reject the calls to halt its development.