This article is about artificial intelligence and machine learning but before with the dive into the nitty-gritty, I would like to tell you a story about my amazing wife Karen. When we got married, we exchanged our vows. The words “love”, “cherish”, “support” were used. (I think I also heard obey 😊). In addition – she probably regrets this decision – but she also promised to make me breakfast every Sunday.
She has been doing it without fail with one or two exceptions. She is a wonderful cook and can make anything. Sunday morning is my favorite day and time of the week. We have a counter in the kitchen. I call it “the magic counter”. Every Sunday morning, I sit in a little stool, I mention what I would like for breakfast, wait by the counter, and food magically appears.
What does this have to do with artificial intelligence? – you say. I envision – in a not too distant future – a kitchen for everyone with a magic counter. Not powered by my lovely wife Karen, but powered by the magic of Artificial Intelligence. We currently live in our society dominated by scarcity. We have a limited energy supply, limited food supply and limitations on other resources.
Let’s analyze for a moment what currently needs to happen in order for “the universal magic counter” to become a reality:
- A farmer must grow the food (planting the seeds, irrigating, harvesting, feed their animals, slaughter their cows, etc.)
- Someone else must transport the food to a distribution center, the supermarket and eventually someone drives it to our kitchen. Maybe us, but more and more Amazon.
- Another human prepares the food.
- Multiply this process many times over by the items that went into the omelet. Eggs, bacon, tomato, cheese, etc.
- Not to mention the gas needed to cook it, utensils, the stool, etc.
Imagine how many people were required to perform the steps above. It involves hundreds if not thousands of people. Now imagine the same process, where many of the people involved in the process are replaced by robots, self-driving cars for transport, automatic tractors, etc. With our current technology and what is likely coming, it’s not too hard to imagine, right? We already have a lot of technology. Now, I probably know what you are thinking – Who’s going to buy all this if people don’t have jobs because the robots took them. Think about this – there have been many technological breakthroughs that have increased our productivity. Locomotives, cars, computers, cell phones, automated manufacturing to name a few. Every time these inventions have come along, the media has whipped us into a frenzy about how we are all doomed and every single time these innovations increase our productivity and have resulted in better jobs and a higher standard of living. But wait you say, “this time it’s different, there won’t be any jobs to replace the current jobs”. I agree that at some point in the future robots will be capable of performing every single job that humans are currently performing.
As that happens, prices, in my opinion, will start decreasing. The cost of delivery will come down because fewer middlemen need to be paid to get the finished product to the table. Couple the removal of humans from the delivery process with a breakthrough in energy technology where electricity becomes much cheaper than today and the cost of food production, delivery and preparation becomes negligible. Fusion technology has broken our hearts many times in the past before, but it still seems like a promising path to cheap energy. Or perhaps wind and solar continue their steady downward path in price until it becomes close to free.
I can envision at least three scenarios as we get to zero-cost resources:
- Everyone forgets because we have lived in relative peace for the last few decades but Russia, China, the US, and a few other countries still have a nuclear arsenal capable of blowing up the world a couple of times over. If some crazy decides to blow us all up, we won’t live to see the end of the story.
- Jeff Bezos continues his relentless march to own every single resource in the world and we all starve because he wants us to. He has an army of killer robots that guard his castle anytime anyone gets close to it and eventually the killer robots become too smart and turn on him and kill him.
- Since resources are no longer scarce, we can all live for free or close to free because resources are now plentiful and easy to produce. Everyone has a “magic counter” in their home (probably powered by Alexa)
There are many other possible combinations and possible outcomes. Unfortunately, the number of scenarios where things go awry versus the number of “happy path” scenarios. Many bright minds are currently trying to come up with the probabilities of bad outcomes. Also, if a person or a group gets a hold of a powerful enough AI, it might only take a few folks to make the decision that they want to wipe out the world. It might not even be a human deciding to end the human race. If an AI is smart enough but feels threatened by humans, they might decide to pull the trigger.
I agree to some extent with Andrew Ng and some of our technology is not evolved enough. But some of the capabilities are already here.
“Fearing a rise of killer robots is like worrying about overpopulation on Mars”
Andrew Ng
Here are some examples of technologies that are already here whether we like it or not:
- Slaughterbots – This term was popularized by an arms control advocacy group. A video was released by this group on YouTube by the Future of Life Institute and Stuart Russell, a professor of computer science at Berkeley in November 2017. The US and Chinese military have the technology to produce autonomous drones armed with tactical weapons that could target anyone in the world by having just their picture. As an example, Chinese military contractors have already started to sell dangerous, autonomous killer robots to customers in the Middle East. For example, the Chinese company Ziyan sells its Blowfish A3 — an autonomous helicopter-like drone armed with a machine gun — to other governments around the world according to Defense One.
- Nanobots – Militaries around the are feverishly working to miniaturize drones to get the to the size of a fly or smaller and use them to target humans. The company Flir has developed the “Black Hornet Personal Reconnaissance System”. The Black Hornet is a minute combat-proven nano-drone which weighs 32 grams and possesses navigation capabilities that don’t require GPS coverage. The drone has advanced image processing and can fly for two kilometers at more than 21 kilometers an hour while carrying a thermal micro-camera.
- Surveillance technology – This technology is becoming pervasive and in the wrong hands, extremely dangerous. On the flip side, in the right hands, it can become a great tool in the fight against crime and other problems. As an example, of a potential benign use, China is currently using thermal cameras to more proactively detect people infected with the coronavirus.
With the advent of nuclear weapons, a select group of individuals has had the capacity to wipe humans from the face of the earth for decades now. The current state of AI can certainly cause heavy casualties and catastrophes but maybe not necessarily wipe out a whole country or planet. Assuming technology continues to evolve and accelerate at the current pace, we are still a few years or decades away from AI having the ability to annihilate every human or conversely completely ending hunger. I do think it is time to start spending time thinking about the possibilities so that we can all enjoy “the magic counter”.
Alberto Artasanchez is the author of “Artificial Intelligence with Python (Second Edition)”