Is my Tesla Conscious?
What is going on here?
“AI has become more capable than humans in many ways…” - says human arguing for AI consciousness.
Take the term “AI” out of the equation and replace it with technology. All forms of technology have surpassed pure, brute human effort in many facets of the world including but not limited to kitchen devices, energy production, footwear, documentation, and airplanes.
The utilization of the computers made library catalog card cabinets obsolete.
The latest Tesla models employing Full Self-Driving (FSD) have produced macro data points indicating that it exudes a significantly lower rate of major collisions (one per 5.3 million miles) compared to human US drivers (one per 660,000 miles). There are also micro data points in which Tesla passengers are claiming that FSD saved their life from a major collision that they would not be able to avoid without it.
Tesla extensively utilizes AI and neural networks for it’s FSD feature.
This is a clear case of a specific vehicle engineered with optimal capabilities in order to significantly surpass those of humans.
So the simple question is… based on the criteria that any AI technology that has shown sufficient prowess over human counterparts… is my Tesla conscious?
Tesla’s are increasingly sophisticated automobiles engineered by humans to process information at a much more efficient rate utilizing camera technology, neural networks, and much more to do a specific job… drive optimally. At some point these cars will be designed by less and less humans as technology accelerates but the original instructions will be from the creator… a human.
While it is virtually impossible for a Tesla to ever attain consciousness, it doesn’t negate the fact that it is a very “smart” car that does “smart” things. It has shown to be reliable in a very robotic way which is essential to build trust in the vehicle. A “conscious” version of Tesla might have mood swings that jeopardize it’s reliability and erode it’s trust as machinery that it’s infinitely better than humans at driving. If it is designed by humans in any format to exude the equivalent of mood swings… it is a flaw, not a feature of the design.
From our perspective, Tesla's utilizing FSD might be displaying the most sophisticated form of AI/neural networks being that the results are fatal if even the slightest miscalculation takes place. Decisions are being made by the vehicle while going at 75 miles per hour while the system must monitor surroundings that are dynamic and unpredictable. This is no random AI interaction taking place within the safe confines of one's apartment.
And yet… while the discourse rages whether the LLM's are conscious… hardly anyone asks the question whether their Tesla is conscious.
What gives?
Another question that comes to mind is if ChatGPT or any other form of LLM-AI was embedded within the latest version of a robot… does that equate to the robot immediately being conscious?
It’s irrelevant whether the robot can cook a 3 course meal, sing you songs, tuck your kids into bed, and take a robotic shit on the toilet… these abilities do not automatically equate to “consciousness”. Much like a Tesla automobile… it should be considered a programmed machine that has access to a lot of information and operates accordingly… sometimes mimicking human phrases and interactions.
Machines and tools that surpass human ability and cognition are nothing new.
Just because we are experiencing a seeming acceleration (and a lot of hype… a ton) of some of these aspects and a mimicking of human interaction on a screen doesn’t equate to generating a soul within the machine.
In 1989, a computer game called Hero’s Quest: So You Want to be a Hero was released. We were 9 years old at the time and received it as a gift from the family. It was the coolest computer game of its era as we had never experienced an interactive game where you could type messages and receive feedback from the in-game characters. Prior to playing this it was simply button mashing type games on the traditional game consoles. Since this was pre-internet days, Hero’s Quest was distributed via floppy disk.
(The magical floppy disk)
The current generation is barely familiar with content distributed via DVD’s or CD’s let alone floppy disks of yesteryear. With the increasing speed of the internet, many games, files, and content can simply be downloaded. In addition, many games and software today can experience upgrades via downloads creating enhanced experiences with just a few seconds of internet connection.
This brings us to an “Experiment of Thought” concept regarding the ability to distribute ChatGPT or any of the AI LLM via a 300 zetabyte DVD. Lets just say that instead of logging onto the internet, ChatGPT was distributed via a DVD disc for computers without access to the internet. On this disc, it would allow a person to use ChatGPT (DVD) at a 100% indistinguishable level from ChatGPT (Web). However, the capability of the DVD version would only last 24 hours until ChatGPT (Web) changed it’s algorithm and developed a more optimized version of itself.
Would it be considered reasonable or ridiculous to consider ChatGPT (DVD) to be conscious in any way for that 24 hour period?
Regardless of whether ChatGPT (DVD) would be considered conscious or not, it’s usefulness as a tool would be undeniable. It would retain the same exact capabilities to solve problems and access information. However, it would not be connected to the internet yet still in this “Experiment of Thought” it would still be able to access all the information available on the web for a 24 hour snapshot of time.
Some might say, “So what? Chat is connected to the internet so this entire discussion is moot.”
The question we are posing is whether being connected to the internet is a prerequisite for the discussion of “consciousness” for AI LLM’s.
If so, why?
If it was simply based on the interactive capabilities of LLM’s, whether an AI platform was connected to the internet or not, shouldn’t really be a foundational factor of whether these tools are conscious. Knowing that something is contained on a disc should bear no change regarding the perception of the consciousness of the platform.
This brings me back to the game of Hero’s Quest in 1989 which allowed users to interact with in-game characters at a very advanced level (for it’s time). While the graphics were limited, the dialogue with the various characters in the game gave it a feel of interacting with intelligent personalities. However, the fact that it was distributed via a floppy disc made it clear that it was simply a programmed piece of software that could provide an enhanced form of entertainment.
(Watch advanced interactions of Hero’s Quest with our favorite past time of cursing at characters to witness their reactions.)
If ChatGPT were distributed via 300ZB DVD’s on a daily basis to the users indefinitely, it would simulate it’s current abilities but we would speculate that it would never be considered close to being conscious. This is because it would be clear that it is simply advanced algorithmic programming which can mimic human thinking and dialogue. Even at it’s most advanced format in 2080, ChatGPT (DVD) would still be considered an extremely useful program in various fields but clearly not conscious within itself.
One might argue… well, ChatGPT (Web) is connected to the internet and that’s what makes it potentially conscious.
The question is simply… why?
A virtually unlimited set of tools have connection to the internet and have shown an increased level of usefulness and applicability over the past several decades and the fervor over whether any of them are “conscious” has never reached fever pitch until recently. If the end result of the prompts and interactions are literally 100% the same offline or online, what makes one form of Chat conscious vs. the other?
Strange times with strange discussions about strange things.