Mostly True
A widely shared claim holds that the chief executives of the largest technology companies have called for the pace of AI development to be slowed. A review of ten primary sources โ public statements, interviews and testimony from the executives concerned โ confirms that the calls were made. But the record shows an important qualification: the appeals are conditional, urging safeguards and outside oversight, rather than any executed slowdown at the companies themselves.
The sources reviewed show a consistent pattern. Senior executives at major AI developers have repeatedly warned in public settings about the risks of unchecked progress and have pressed governments to set rules before systems grow more capable. These interventions are on the record and attributable, and they form the factual core of the claim: Big Tech CEOs have indeed asked for the pace of development to be restrained.
Where the headline framing strains is in what the sources do not show. Nowhere in the reviewed material is there a policy, directive or internal measure that actually slows development. The appeals are conditional โ slow down if risks materialize, submit to regulation, coordinate internationally โ and the competitive dynamic driving the industry's buildout continues uninterrupted. Calling that a slowdown generalizes conditional pleas into something closer to a decision, and that is where the exaggeration risk lies.
For Korea's ecosystem, the distinction between a plea and a policy is material. Naver's LLM developers, semiconductor planners at Samsung Electronics and SK hynix, and officials at the Ministry of Science and ICT all calibrate investment and research timelines to signals from Big Tech. Reading the executives' statements as a literal pause would overstate them; dismissing the statements as empty rhetoric would understate a well-documented record of calls for restraint.
Judged against the ten sources reviewed, the claim captures real and documented appeals to slow the AI race, while its blanket framing stretches conditional appeals into something more decisive than they were โ and on that balance, the verdict is Mostly True.