December 6, 2023

Machines tend to take things literally. As far as your computer is concerned, the world is made of 0s and 1s and governed by algorithms. Even artificial intelligence sees and comprehends the world through a purely computational lens: ChatGPT-4 is many things — frightening, groundbreaking, encyclopedic, sometimes wrong — but it isn’t funny.

Then again, humans don’t expect AI to be funny. Sure, Siri has a habit of responding to questions about its love life with evasive and often amusing answers, but we all know some programmer at Apple came up with them. We like to think it requires wit and wisdom to crack a joke, and we prefer to believe those are qualities we possess that machines can’t. This weekend, however, that assumption came to an end: One AI bot may have just figured out sarcasm.

Musk’s Grok

After just two months of testing, Elon Musk’s newest pet project, xAI, announced the release of its first chatbot: Grok. Made in the image and likeness of its creator, Grok is more than a digital nerd with an encyclopedic ability to generate Wikipedia-type answers — it’s supposed to possess a flair for sarcasm, a “rebellious streak,” and a constant interest in the nonsensical conversations going on at X.

According to a statement released by xAI, the creation of Grok was inspired by Douglas Adams’ Hitchhiker’s Guide to the Galaxy — a book Musk has talked about in interviews as having had a profound influence on him when he was going through an existential crisis as a teenager.

The chatbot’s name, however, comes from another sci-fi novel: Robert A. Heinlein’s Stranger in a Strange Land. In that book, the term “grok” originates with the Martians and doesn’t have an English equivalent (although, according to Merriam-Webster, it’s also a verb meaning to “understand profoundly and intuitively”).

Whatever its origins, Grok seems to have the promised sense of humor. On Friday, Musk published a response the bot gave when asked for a recipe for cocaine:

Grok also has thoughts about Sam Bankman-Fried’s ongoing court battles:

Not only does the bot possess a sense of humor, but it also has no cutoff date when it comes to resources. While ChatGPT is still living in January 2022, Grok is constantly inundated with the latest on X — a tactic that, while unlikely to result in 100 percent accuracy, could make it a source for Americans already treating social media as a newspaper. (READ MORE from Aubrey Gulick: Biden Is Finally Regulating AI … for Equity?)

The chatbot is certainly still in its infancy. Its Sunday release, which constituted the next phase of the testing process, was restricted to a small number of beta users (there’s a waitlist). When it is released to the public at large, it will be available to users on X who are willing to pay the $16 per month for X Premium.

AI Is Risky

Musk’s announcement comes days after global public figures from 28 countries, including Musk, attended the AI Safety Summit in Bletchley Park in the U.K. and agreed to a global declaration on the need for “safe and responsible development of frontier AI.”

During the summit, Musk warned that AI carries some inherent risks: “On balance, I think AI will be a force for good most likely, but the probability of it going bad is not zero percent so we just need to mitigate the downside potential.” He added, “I think this is 80% likely to be good, and 20% bad, and I think if we’re cognisant and careful about the bad part, on balance actually it will be the future that we want.” (READ MORE: Physiognomy Is Real, and AI Is Here to Prove It)

In Grok’s case, the worst it will likely bestow on the world is an inundation of sarcastic quips. Sarcasm is, after all, the lowest form of humor, so it’s no wonder it’s the first AI acquired.