• TheTechOasis
  • Posts
  • Palantir just showed off an AI that can go to war

Palantir just showed off an AI that can go to war

Peter Thiel might have just shown us the future of warfare

🏝 TheTechOasis 🏝

🤖 This Week’s AI Insight 🤖

ChatGPT, Your Country's Brand New War Weapon

If I were to recommend a TV series, I would probably recommend The Blacklist, a series about a special FBI task force led by an infamous criminal, the great Raymond Reddington.

In this series, one of the main characters is Aram Mojtabai, a genius hacker.

He’s always portrayed as doing crazy stuff with a computer, something that’s nothing more than a Hollywoodian, glorified version of what using military-grade software really is.

Or is it?

Because with the release of AIP, Palantir takes the talents of ChatGPT and deploys them into full-out war.

Now, military operators will feel like Aram… or even better.

Source: Author with MidJourney

War requires agility

While consumer-end technology is usually focused on reducing friction and making our lives easier, when it comes to war is all about raw power and quick decision-making.

In fact, military technology spearheads many of the most advanced technologies in the world.

To be specific, defense budgets have all but decreased, reaching $2 trillion in 2022 according to the IISS.

Naturally, it was a matter of time until Generative AI was used for war.

And that time is now, as Palantir, a leading US government contractor, has announced the AIP, its Artificial Intelligence Platform.

An LLM-powered stack

The AIP is comprised of a series of powerful capabilities.

  • It can easily integrate any LLM, like GPT, with any classified or non-classified networks, allowing these models to access unique, real-time military data.

  • It allows you to decide what LLMs can or can’t do or access, giving a much-needed sense of control over them

  • Includes a series of capabilities to ensure full traceability of data and actions taken by the models

And Palantir’s goal with AIP?

“Ensure that LLMs are used in an ethical and legal way”

Sounds like a marketing stunt more than anything.

But even though we can’t trust the “kind intentions” a for-profit company like Palantir allegedly wields, fully-traceable solutions like AIP are better options than no-guardrails software to ensure and enforce accountability in situations where people, including kids, will most likely perish.

But what really made my jaw drop was the demo they showed.

Fully-automated warfare

In short terms, the main interface of the tool is an AI assistant you can talk with, as you can see in the image below:

Source: Palantir

Sounds like ‘meh’ stuff… until you forget that, unlike your ChatGPT plus, this agent is connected to some of the most advanced military weapons in the world, and can take action for you.

In the demo, the conversation with the agent portrays a military action against a recently-detected enemy.

The AIP agent’s actions include (all based on requests by the human):

  • Alerting of the discovery

  • Deploying an MQ-9 drone to gather additional imagery (image on top)

  • Generating three courses of action based on the constraints of that particular mission

  • Escalate the three options to the commander at the request of the operator

  • Upon the chosen tactic, the GPT agent drafts the plan, including the route and the military capacity of the deployed team

  • Jamming the communications at the request of the commander

  • Finally, deploying the Tactical team to start the mission

All this based on one single conversation that leverages, in the back, all the necessary, up-to-date data, be that classification markings (top-secret, confidential, etc.) or non-confidential.

AI opens a new chapter to worry about

Although expected, the quick deployment we’re seeing of Generative AI technology into warfare is a troubling indication that the world is being driven into AI dominance.

On the flip side, this could be an indication that the future of war is non-human, which would reduce casualties from soldiers.

However, AI could make those that harness its power much more deadly, which could in turn generate far more casualties as a whole.

At the end of the day, Generative AI is all about optimizing the way we communicate and take action with machines.

And that, sadly, includes war.

Check the video of the demo here

👾Top AI news for the week👾

🧐 Should models like ChatGPT be biased? Renowned USC researchers delve into this controversial topic

🖋 AI will be teaching writing skills to kids, says Bill Gates

🕹 You can now control what happens to your data with ChatGPT

😍 Stability AI has released two image upscalers to generate super high-definition images from smaller ones (link includes examples)