X
search

OpenAI signed a contract with the US Department of Defense (but don't worry)

Pictured: Arthur Brown
By Arthur Brown
January 18, 2024
Featured image for OpenAI signed a contract with the US Department of Defense (but don't worry)

Generative AI is still relatively new compared to much of the tech running our world. So, its potential wide-reaching effects are still an enigma. This is why we don’t want it around any major governmental bodies… Anyway, OpenAI just signed a contract with a major governmental body. According to sources, OpenAI signed a contract with the US Department of Defense (DOD).

The OpenAI DOD contract isn’t as scary as it sounds

However, you can stop looking up “How to build a nuclear bunker” on Google. The government isn’t putting ChatGPT in charge of the nation’s military or any big missiles. It’s going to be used for cybersecurity reasons. We don’t know what software the DOD is going to be using, exactly. We’re not sure if it’s going to be using GPT-4 or if OpenAI is going to develop a specialized model for these uses.

Advertisement
Advertisement

In any case, the contract is to help boost the country’s cybersecurity. OpenAI will lend its software to help develop open-source cybersecurity tools. So, no missiles, fortunately. Also, this technology is going to be used to participate in the DARPA AI Cyber Challenge. This is an ongoing competition for computer scientists, AI researchers, developers, ETC. to help develop powerful cybersecurity tools and patch critical vulnerabilities. With OpenAI’s powerful AI tools, we’re sure that it’s going to be a formidable contender.

Should you be worried?

Well, we’re still at the very beginning of AI’s involvement in government affairs, so this remains to be seen. The government is only using these tools to help develop cybersecurity tools, so it won’t be in charge of anything. All of the major decisions will be made by human beings.

However, the issue is that AI technology isn’t 100% reliable just yet. There’s still the chance of AI hallucinations. As much as companies try to reduce this phenomenon, it still happens. That’s not as bad if it hallucinates while generating a recipe or a book report, but it’s terrible if the nation’s cybersecurity is at stake. This is something that we’ll have to wait to see.