Live data from Hacker News

OpenAI drops ban on military tools to partner with The Pentagon

semafor.com

11–20 of 509 posts

Re: OpenAI drops ban on military tools to partner with The Pentagon

#13

[flagged]

Exactly. Every American company should be proud to be a partner to the Pentagon. They don’t choose their missions but overall the world is a better place with American hegemony expressed through her military superiority. It would be a tragedy to lose hold of the world order because an arbitrary board disagrees with how congress uses them.

Re: OpenAI drops ban on military tools to partner with The Pentagon

#14
Nothing says "we're concerned about AI risk and harm to human life" quite like partnering with the military-industrial complex.

It's even better timing to do this while the Pentagon is providing material and intelligence support to a government whose decisonmakers' public statements are explicit admissions of intent to commit war crimes and also convey genocidal intent.

Glad we spent all that time talking about AGI and Roko's Basilisk, those should be top of mind always, never the current actions of the humans in charge at both OpenAI and the US government.

Re: OpenAI drops ban on military tools to partner with The Pentagon

#16
The CIA probably realizes it doesn't need to fund rebel groups [1] or use local journalists/clergy [2] to instigate a regime change these days - they can just flood a target country's social media with AI-generated propaganda.

1. https://en.wikipedia.org/wiki/Iran%E2%80%93Contra_affair

2. https://www.intelligence.senate.gov/sites/default/files/hear...

Re: OpenAI drops ban on military tools to partner with The Pentagon

#17

[flagged]

This is the right take. AI on the battlefield is a foregone conclusion, so it behooves us to understand and implement AI earlier, rather than later. That way we can do it in a measured and tested way; the alterative is realizing our lack of AI is causing us to lose a war and accelerating implementation of it as much as possible with less regard for safety.

Battle AI stuff is already developed in-house or by firms like Palantir.

OpenAI's products are more likely to be used on the billions of emails sent daily from people who aren't in full command of the English language or are too lazy to look at man pages. Aka regular LLM usage for normal people.

Re: OpenAI drops ban on military tools to partner with The Pentagon

#19

The CIA probably realizes it doesn't need to fund rebel groups [1] or use local journalists/clergy [2] to instigate a regime change these days - they can just flood a target country's social media with AI-generated propaganda. 1. https://en.wikipedia.org/wiki/Iran%E2%80%93Contra_affair 2. https://www.intelligence.senate.gov/sites/default/files/hear...

I would wager a whole lot of money that the CIA is very much on top of this one.

Re: OpenAI drops ban on military tools to partner with The Pentagon

#20

[flagged]

> US military must be the ones possessing and in control of the most advanced AI technology.

Except that in this case they aren't possessing or in control of it. They're just writing checks to a third party to use what they possess and control.

I do think that once we have real AI the government should be all over that, but I don't see much need for them getting too worked up over a chatbot with a tendency to outright lie. I'd certainly hope that they don't use OpenAI's questionable product for anything important or anything that would have a meaningful impact on people's lives. I wish that for every company actually, but when it comes to where bombs drop or who drones kill its really important to keep humans involved and to make sure your information is accurate.

Post reply on HN