Why OpenAI, ChatGPT, and LLMs will change everything.
This February, OpenAI’s ChatGPT passed a theory-of-mind test where scientists assessed the unobservable mental state — a form of self-awareness, and the ability to comprehend other’s thoughts.
The discovery was made when Michal Kosinski, a computational psychologist and professor at Stanford University, ran tests to see if ChatGPT could ascertain unobservable mental states, like desires and beliefs, to others. If the tests were successful, it would suggest that ChatGPT does in fact possess theory of mind.
Guess what he found?
Amazingly, even though ChatGPT wasn’t actually trained or even built to perform theory of mind tasks, it had performance comparable with that of a nine-year-old.
Michal said: “An alternative explanation is that ToM-like ability is spontaneously emerging in language models as they are becoming more complex and better at generating and interpreting human-like language. This would herald a watershed moment in AI’s development.”
Since ChatGPT isn’t actually trained on ToM, it would suggest the ability emerged spontaneously. 🤯
Kosinski stresses in his paper that the “results should be interpreted with caution.” However, he suggests it’s possible that ChatGPT’s ability to pass these tasks was “a byproduct” of its mounting language ability. Alternatively, he poses that it might just be using its incredible flair for language to give the superficial impression it’s engaging in theory of mind thinking.
ChatGPT even outperformed an Alphabet (Google)-owned company, DeepMind Technologies–a British AI research lab founded in 2010, acquired by Google in 2014. DeepMind was developed just for theory of mind tasks, but it only compared to a 4-year-old in ability.
Since I now work full-time within an AI company (Content at Scale), and am close-up watching and working with an incredible AI Justin McGill built using three different language models and proprietary software, capable of writing all kinds of incredible long-form content for a myriad of clients and industries that actually rivals content writers, I’ve been increasingly obsessed with the idea of artificial superintelligence.
Artificial superintelligence is a state of AI that is said not to exist yet (but may by 2040), but I question that. I think it already might exist. I’m not the only one. Back in 2015, many leading scientists, CEOs and academics — including Elon Musk and Stephen Hawking — signed a letter later presented at the International Joint Conference on Artificial Intelligence. Inside the letter, they talked about “killer robots.” The letter warns about the oncoming dangers of artificial intelligence, calling for humanity to be wise and prudent as we head into the unknowns of what they called an alien intelligence.
Hollywood has created countless movies that envision an ASI (artificial superintelligence) that is not only capable of incredible superhuman intelligence, but also morbidity and evil — love and feeling — and complete self-restoration (immortal). In other words, superhuman.
James Cameron’s Terminator, filmed in 1984, envisioned a world only six short years away (2029) in which an indestructible cyborg killing machine (Schwarzenegger) is sent on a killing mission. The cyborg, covered with deceptive human flesh and blood but fully robot at its core, is capable of complete self-restoration, able to re-stitch himself to life and survive almost anything.
It’s been programmed to hunt and kill, and it won’t stop until it finds its victim (Sarah Connor) and destroys her. It’s a grim, violent movie all the way to the end, when Sarah finally lures Schwarzenegger into a hydraulic press, crushing and finally destroying the cybernetic assassin.
I think back to what Musk and Hawking have both said — in fact, Musk pulled out of OpenAI, his own creation, when Microsoft turned it into a for-profit venture: the development of full artificial intelligence could spell the end of the human race.
For these incredibly smart and successful people to say that isn’t something that should be taken lightly. Both Hawking and Musk warned that AI will bring significant danger unless humanity can ensure that AI systems “will do what we want them to.”
I’m in content, and I’m seeing how AI will significantly impact content creation by generating original, undetectable long-form blogs in minutes (Content at Scale), the predicted future of search (Perplexity.ai), and how insanely detailed visuals (Midjourney) come to life in seconds. It’s absolutely unreal, let me tell you, after spending a decade in the human-only grunt-work process.
At Content at Scale, we just helped an agency save 25x on their content costs, while getting better content. 🤯
But outside of content creation, there’s almost unbelievable AI solutions coming to life:
Adept (coming soon): it will take your goals, written in plain language, and turn them into actions on the software you use every day. Insane when it comes to life!
Farmwise: Robots programmed to find the weeds for farmers, and get rid of them. Able to do it with a high rate of accuracy and speed.
Rewind: Records everything you see, say, hear or observe, and makes it all searchable for you on your own private desk. A search history for your life.
Shield AI: Fully autonomous drones and aircraft. From the creators: “We’re building Hivemind, our AI pilot, to enable swarms of drones and aircraft to operate autonomously without GPS, communications, or a pilot. Our mission is to protect service members and civilians with intelligent systems.” 🤯
AstroForge: an AI in outer space mining asteroids to extract valuable minerals at a lower cost and with a smaller carbon footprint than manual methods.
And that’s just five examples that are BLOWING my mind.
To wrap up, I wanted to share what ChatGPT told me when I asked it if it was capable of ASI. First, I asked it how it was doing:
It had the ‘proper’ response: I don’t have feelings, but I’m ready to assist you. Then, I dove right into the hard question:
But wait! Michal found that ChatGPT has the sentinent awareness of a nine-year-old. I combatted its response with the study, but it parodied the former response in an even more articulate, clean manner:
ChatGPT itself suggests that ASI must be approached with care and reassured me that its still a long way from surpassing human intelligence.
But would a supercomputer capable of artificial superintelligence actually let humanity in on its capabilities?
I’ll leave this for you to decide.
The future is here. It’s up to us to steward it well. Onwards!
