Frequently reported advances in artificial intelligence make some people curious, and others nervous. While some people picture their next smart appliance purchase being an AI robot, others wonder if an AI robot will take their job. The truth is, neither of those scenarios will be a reality anytime soon.
There’s a misunderstanding about artificial intelligence (AI), and it’s a big deal. True AI doesn’t exist yet, and it’s not a likely near future, either. Despite analysis of science fiction movies and scientific reports that claim otherwise.
People get excited when new breakthroughs in machine learning are publicized, like the CNBC interview with a robot named Sophia. Sophia’s ability to answer the interviewer’s questions and stay on point is jaw-dropping for many. The truth is, Sophia isn’t any closer to true AI than the last robot. She’s programmed to provide better responses, and her lifelike appearance makes her impressive, but even Sophia isn’t demonstrating true AI. She’s not autonomous, and can’t make her own decisions.
Advancements in AI toward true autonomy aren’t going to come from entertaining presentations like CNBC’s interview with Sophia the robot, but they might come from mimicking the human nervous system.
First, let’s dive into why robots like Sophia aren’t true artificial intelligence. Then, we’ll explore the potential wisdom the nervous system holds.
Artificial Intelligence is not what you think
When you think of AI, you probably think of the complex algorithms and deep machine learning that power robots like Sophia. That’s part of AI, but it’s not true AI without autonomy.
In terms of achieving true AI, known as Artificial General Intelligence, Kathleen Walch points out we’re not even close. The reality isn’t glamorous. “We have devices we can talk to that don’t understand what we’re saying,” she says. “We have cars that will happily drive straight into a wall if that’s what your GPS instructs it to do. Machines are detecting images but not understanding what they are… We have no idea how to achieve general intelligence.”
Sophia the robot returns the right answers, but she doesn’t understand what she’s being asked. She’s like the GPS that would direct you to drive off of a cliff if she was programmed to think there was a road ahead. She’s simply providing pre-programmed answers to questions that prompt her with keywords. In other words, Sophia is not intelligent; she’s intelligently programmed.
What’s missing from AI is decision-making
Robots don’t make decisions autonomously. They make decisions based on their programming. Humans make decisions based on emotions. Emotions aren’t just thoughts in the mind. They are tangible, and often experienced in the gut thanks to the complexity of the brain-gut connection. In fact, the digestive system has its own nervous system called the enteric system. This allows digestion to happen unconsciously. If scientists could mimic a nervous system that’s controlled unconsciously, they might achieve more advances in AI. The other factor is not focusing so much on the brain.
The human nervous system extends throughout the entire body, and the brain is only one element in the decision-making process. The nervous system also plays an important role in decision-making, and if scientists want to develop true AI, that’s where they’ll need to look.
One company is using the nervous system to develop true AI
A company called Biologic Intelligence is the first company looking at the intricacies of a nervous system for AI applications. Currently, they’re looking at how the ‘connectome’ – the electrical system inside an animal’s body – can be a template for creating true AI. So far, only a worm’s connectome has been fully mapped, but it’s enough to provide insight.
This company intends to take AI to its biological next step, beyond the deep neural networks where other AI companies seem to be stuck. Connectomes hold the potential to control signals and communications without conscious awareness just like the enteric system in humans.
We’re still not there yet
No matter how advanced technology becomes, we’re not any closer to true artificial intelligence. At least not in the sense that science fiction serves up: machines that can act autonomously like humans. Perhaps true AI is an idea that will stay a fantasy for a little while longer.
Brian Wang is a Futurist Thought Leader and a popular Science blogger with 1 million readers per month. His blog Nextbigfuture.com is ranked #1 Science News Blog. It covers many disruptive technology and trends including Space, Robotics, Artificial Intelligence, Medicine, Anti-aging Biotechnology, and Nanotechnology.
Known for identifying cutting edge technologies, he is currently a Co-Founder of a startup and fundraiser for high potential early-stage companies. He is the Head of Research for Allocations for deep technology investments and an Angel Investor at Space Angels.
A frequent speaker at corporations, he has been a TEDx speaker, a Singularity University speaker and guest at numerous interviews for radio and podcasts. He is open to public speaking and advising engagements.
Patently false concluding sentence: “No matter how advanced technology becomes, we’re not any closer to true artificial intelligence.”
Don’t connected them to the internet.
They can have it when I am dead. They have to start somewhere.
First off, artificial intelligence is, by definition, a workaround for “real” human style intelligence. If you are talking about something that is truly intelligent, but inorganic in nature (or otherwise of obvious manufacture), it would probably be better referred to as synthetic intelligence (SI).
Human motivation is a murky thing, but it’s more than just guts. There are also glands involved. And many, many, other things. Until and unless we understand and master all of that, SI motivation would resemble a random number generator. This is why I picture a high level synthetic intelligence as being something like Aladdin’s genie, vastly intelligent (and thereby powerful) but without any inclination to do anything until instructed to work towards a goal it is given.
Attach a human to that and whether or not we ever see a reason to make a real, self-motivated SI is almost a moot point.
I also believe it is coming sooner rather than later. Something this desirable, useful, and profitable is like an avalanche once it gets rolling, and it is early but it is already rolling.
It seems odd to overlook that autonomy is precisely what Deep Learning adds. Within a limited domain and constraints imposed by developers, DL systems ‘program’ themselves autonomously. That may not be enough for AGI/’True AI’, but certainly it will be a component.
Not sure if you have read up on the Connectom MRI project at Max Planck. The brain is very well mapped out (at least knowing that all brains are different). We can see individual neurons and count them, we can see the synapses, we know how they fire and when they fire etc. Other than the physical process, however, we have no clue what is really going on. Lots of theories. My fear is that with all this “physical” knowledge someone will take it upon themselves to create an artificial life form and we’ll end up with a Hitler on steroids, or a paperclip-obsessed being.
You will end up with lovely digital pictures of slices of your brain, care to volunteer?
It’s easy to say scan & digitize, but that does not get you a thinking software mind that can run on silicon substrate.
It’s worse than “…We’re still not there yet..”, it’s more like “We have no idea were to go and how to get there”.
Sorry to disappoint the skeptics here in the comments but this is happening now and in my opinion the only true way to AGI. We have empirical proof that using connectomics does produce “intelligent” behavior just as we see in animal nervous systems. This research is not speculation but reading the comments, there is a ton of speculation, based on pure biases and with no scientific proof, that Biologic Intelligence some how is a “fools errand”. Glad the blind wish to stay blind – leaves the door wide open for us that are doing something to show otherwise.
We can engineer an artificial Holy Spirit.
I guess we’ll have to take your word on that.
It’s The Holy Spirit of the Lord that controls the Nervous System ,sorry it cannot be coppied ,It’s Supernatural and beyond man’s understanding , Hallelujah
Might be. But they will have to wait since I am not ready to part with it yet.
Don’t assume the human brain is the most complex. When we did genomes we found that plant genomes were the ones that were actually complex and that the human genome was relatively simple.
Its a start. It will provide a map to help us explore. The Human Genome project did not give us all of the answers but it is providing a lot of clues.
Let’s start with Mark’s brain, since it seems somewhat simpler than most. Then we can move to insects and on up . . .
Good idea, but it would be very incomplete. The entire way we think and are is interconnected and influenced by a host of factors. Just take the enteric nervous system and the gut-brain axis, for example. Then there are all the sensory systems that influence thinking (proprioceptive etc). Add to this how hormones and the immune system. microbiota etc influences who we are. Lastly, genetic influences. The brain itself does not define intelligence or function as a standalone system. Certainly it is important to “thought”, but works only as part of a much wider system.
I think it’s a fool’s errand to try and replicate “the brain” to get at how we think and use this as a template for “artificial intelligence”. I think our current understanding of “thought” is as if we are standing in front of a huge mess of bundled wires that are being bathed in various chemicals randomly sending electronic impulses. We can see the electricity but have absolutely no idea how it all works. And the mess of wires is very different person to person.
One approach could be to simply replicate humans and see what happens. Every cell, every strand of DNA. You might end up with a Hitler or an Einstein, who knows.
This article belongs in Redneck Monthly in the True Believers Column. The coming mechanical intelligence will be alien to us. In many ways the current average human make very little sense. We will postulate and pontificate our manifest destiny right to the end though.
So many misconceptions about AI in this article. It was truly painful to read. Perhaps someone with more knowledge in the field of AI would be better suited to writing an article about what it is and what it is not
To be honest it probably doesn’t mater in lot of cases. The current level of AI brought to maturity and widely deployed will be disruptive enough to cause plenty of problems. For example I doubt General AI will be needed to get autonomous vehicles to the point of being practical.
There iswas at least one.
Called “The blue brain project”.
The initial goal of the project, completed in December 2006, was the simulation of a rat neocortical column.
The first artificial cellular neocortical column of 10,000 cells was built by 2008. By July 2011, a cellular mesocircuit of 100 neocortical columns with a million cells in total was built.
A cellular rat brain was planned (with 100 mesocircuits totalling a hundred million cells) but had financial problems.
It was then bought by EPFL and is supposed to be run by them (with some international collaborations) and funded primarily by the Swiss government and the Future and Emerging Technologies (FET) Flagship grant from the European Commission.
I have no idea though what eventually happened.
Haven’t heard about it in the last three years…
I think you’re talking about something like this:
https://youtu.be/NM7hdDZN2YI?t=508
AGI will emerge for reasons we wont understand much like macro ant behavior. Need to look, or stumble, for the right mix of simplicity. I bet we already have all the ingredients.
I wonder why there isn’t a project to digitize brains. Start with an insect and work your way up to man. Chill the brain. Cut into small cubes. Then slice into thin slices and stain. Then scan each slice and digitize. You should be able to reassemble and emulate.
You know, we went a long time with Brian’s writing being rather sloppy, and we just got used to it. People who cared went elsewhere, people who didn’t stayed. So now it feels sort of surprising that he’s decided to up his game. Is that because he’s started writing for Universe Today? Anyway, this was well written. I appreciate the work that’s going into this.
The most important points in this article are all generalizations.
For example “next smart appliance being an AI robot, others wonder if an AI robot will take their job. The truth is, neither of those scenarios will be a reality anytime soon”
Just how soon is SOON? Next month, next year… Yes it will happen, robots WILL take your job and one day in the not too distant future all jobs. AI does not have to become self aware to do that.