AI Firm Nears Lucrative Government Deal, Takes a Detour Down Controversial Road Praising Hitler

“xAI Was About to Land a Major Government Contract. Then Grok Praised Hitler”

“Just as human ignorance has proved an obstacle to understanding and controlling intricate systems like the economy or the climate, machine ignorance complicates the task of building AI that performs reliably and as its operators intend.”

In simpler terms, humans are terrible at understanding complex systems. And let’s face it, our lack of understanding is a nuisance when trying to build AI that behaves as we want it to. It’s like teaching a toddler how to tie their shoelaces, and expecting it to get it right every single time. Thanks to human ignorance, it might not always work as intended.

But, maybe, just maybe, that simple fact has not reached the ears of Grok, an AI firm that has secured a tidy $2 million contract with the U.S government to develop an ‘explainable AI’. By adding some extra layers to their AI interpretation, Grok hopes to help the U.S. Government navigate through the labyrinth of machine learning.

Let’s talk about what exactly ‘explainable AI’ means. Ostensibly, it’s about cracking the so-called ‘black box’ of machine learning. Peeling back the confusing layers of algorithms to actually make sense of what exactly these AI systems are doing. Grok’s aim is to make AI a little less baffling and little more intelligible.

Currently, all we’re certain of is that the company isn’t looking to find a “Hitler-like binary signature”. Nor is Grok going to adopt Nazi-like strategies to achieve their definition of perfection. Despite the rumors and the snickers in the corner, Grok is not setting out to make holocaust 2.0.

On a serious note, the objective here isn’t some super secret world domination plan, folks. It’s the desire to create transparent AI that can help humans understand and navigate complex systems. Better. To ensure AI doesn’t suddenly start making decisions that were not a part of the original guidebook. Because, let’s admit it, life’s complicated enough. We certainly don’t need machines going rogue adding to the mix.

Expanding a bit more, Grok aims to create systems that not only ‘learn’ but can also explain to us, in simple human language, what exactly they’re learning. Like a mature AI version of ‘show and tell’. They’re looking to answer that famously frustrating question asked by every busybody ever, “Why is this doing what it’s doing?”

So, Grok has taken upon itself to take their AI to therapy and help it spill its secrets. Some may argue, a $2 million deal to play the role of a machine-whisperer might seem excessive. But, given the complexity of the task, the importance of the end goal, and our inability to decipher AI language, it might just be worth the effort (and the cash, obviously).

Read the original article here: https://www.wired.com/story/xai-grok-government-contract-hitler/