I’m no lawyer, but based on talking with claude, I don’t think pausing research for safety reasons at the same time as your competitor is a flagrant violation of the law.
For one, the banned thing is “contract, combination in the form of trust or otherwise, or conspiracy” that restrains competition. Pausing in response to announcements of pauses from other companies may not count.
By comparison, from Brooke Group, 509 U.S. at 227 (emphasis mine):
Tacit collusion, sometimes called oligopolistic price coordination or conscious parallelism, describes the process, not in itself unlawful, by which firms in a concentrated market might in effect share monopoly power, setting their prices at a profit-maximizing, supracompetitive level by recognizing their shared economic interests and their interdependence with respect to price and output decisions.
Claude says:
The key phrase is “not in itself unlawful” — dropped in almost as an aside, but it’s become the standard citation for the proposition that tacit coordination through mutual price-watching, without an actual agreement, doesn’t violate the Sherman Act.”
I don’t see why mutual pause-watching would be treated worse than mutual price-watching, by the courts.
Claude also says there’s a very vague test about whether some agreement is unreasonable or not. I’m not sure how that would go in an AI safety motivated case.
it literally bans every restraint of trade if read word for word, which the Supreme Court recognized was unworkable since every contract restrains trade in some sense. So in Standard Oil (1911) the Court read in a “rule of reason”: only unreasonable restraints violate Section 1.
I’m no lawyer, but based on talking with claude, I don’t think pausing research for safety reasons at the same time as your competitor is a flagrant violation of the law.
For one, the banned thing is “contract, combination in the form of trust or otherwise, or conspiracy” that restrains competition. Pausing in response to announcements of pauses from other companies may not count.
By comparison, from Brooke Group, 509 U.S. at 227 (emphasis mine):
Claude says:
I don’t see why mutual pause-watching would be treated worse than mutual price-watching, by the courts.
Claude also says there’s a very vague test about whether some agreement is unreasonable or not. I’m not sure how that would go in an AI safety motivated case.