Which side of the current disagreement over AI military usage do you support more: Anthropic or the US Defense Department?

Anthropic (78%) – "I get it, no consumer wants to be told how they can and can't use the product they're paying for after they've bought it. But AI isn' a pharmaceutical product or vehicle. Anthropic clearly knows its product has the potential to either be abused or flawed enough to be harmful. I have a lot of respect for them risking their contract to protect the American people, dare I call it an act a patriotism?"

• "Until more testing and analysis is completed, using unrestricted AI for unmanaged combat is opening pandora's box. AI technology is too early in its development stages. Its use for surveillance is a violation of our individual rights."

"It sounds as if Anthropic, who understands AI better than anyone, feels there should be some guardrails. I trust them to see the need for those guardrails vs. the government (any government republican/democrat). We are entering unchartered territory, and we need to be able to trust those who know. If a company is willing to give up millions of dollars for their concerns, it seems like those concerns are well deserved."

• "Those guardrails strike me as both moral imperatives and incredibly reasonable. Which is nice! Sometimes there's not much overlap in that venn diagram, but here it's more of a circle. The Pentagon trying to get around that, at least from the outside, and particularly the mass surveillance bit...that's ominous."

"AI systems shouldn't be making life or death decisions autonomously. Besides the doomsday parallel, when it isn't, it's massive government overreach. Combine that capability with mass surveillance and you have a government that can make a robot to hunt anyone that's tracked that it labels an enemy of the state; then without trial, jury, or judge, execute that person. A human can refuse to do something like that. A human can refuse to break the law. AI has shown that despite its ability to know all the text in the law, it isn't good about actually following it, even when not prompted to directly break it."

• "Non-human decision making lethal attacks is terrifying because who suffers the recourse of a mistake? The facial recognition stuff is scary, but pretty sure it already exists at a high competency so not so bad. If you're in public, you're open to being recorded and identified."

"Government reach is already out of control. Drawing a line in the sand at AI assisted mass surveillance of Americans and indiscriminate war methods is totally reasonable and “anthropic”. Good for them for keeping to their mission statement and standing up for their boundaries. Finally, a company that cares more about humanity than money. Sadly, I’m sure the next AI company employed by our military, if Anthropic is ousted, will gladly throw American privacy and civilized warfare to the wayside. The hoax of mass surveillance for “safety” and “because China is a threat” is a ruse that will be used again and again until control and surveillance of Americans is indiscriminate. Anthropic is standing up for Americans and humanity, and they should be supported for that stance."

• "I can't believe I am agreeing with an AI company but there is a first time for everything. If this company doesn't want the US government using their system to spy on american citizens then that's their perogative. If the Pentagon doesn't like the limits put them by the companies who own this software then they should design their own not bully companies into complying with their demands"

US military (15%) – "It is the best solution for America’s national defense as hostile situations develop in mere minutes in today’s world."

• "In peacetime they are the most bogged down inefficient bunch due to regulation and large organization handicaps and stupidity. They do seem to work well in wartime scenarios as they toss the regulations aside to efficiently win the war. Anything they do using AI needs to be overseen by independent CLEARED teams using concrete guidelines. 'Lawful' is a good guideline to start with for citizens of the US, not foreign adversaries and enemies which ignore laws anyway."

Unsure/other (7%) – "This important matter should be vetted through congress! Not a rouge DOD."