A year after the clash of the tech titans, are we any closer to facing Elon Musk’s dire AI Apocalypse predictions?
Controversial or not, I support Elon Musk’s growing contributions to technology, his drive to reduce climate change and his powerful entrepreneurial spirit. As for his numerous new projects like the Hyperloop, Musk electric jet and Solar City, I’m cheering them on as well.
But I don’t think he’s right when it comes to AI predictions. About a year ago, Tesla and SpaceX CEO Musk locked horns with tech titan and Facebook CEO Mark Zuckerberg on the possibility of an AI Apocalypse. This scenario, which Musk views as very plausible, envisions robots with the power to take over the world and possibly even destroy their human programmers.
For years, he’s warned about doomsday scenarios ranging from a world where humans are ruled by devices with artificial intelligence to an actual Terminator-like robot uprising. His warnings are calamitous. “I keep sounding the alarm bell, but until people see robots going down the street killing people, they don’t know how to react, because it seems so ethereal. Mark my words – AI is far more dangerous than nukes,” Musk said. He even founded nonprofit OpenAI in 2015 to develop artificial general intelligence “in a way that is safe and beneficial to humanity.”
Last summer, Musk asked a group of US governors at the US National Governors Association summer meeting for proactive regulation of artificial intelligence because he believed that AI posed a “fundamental risk to the existence of civilization.”
Musk’s push for action was met with rebuke by Zuckerberg, who responded through a Facebook Live broadcast that his doomsday AI scenarios are unnecessary and “pretty irresponsible…. I am optimistic. And I think people who are naysayers and try to drum up these doomsday scenarios – I just, I don’t understand it.” Instead, Zuckerberg foresees a world where AI will help save lives through self-driving cars and greater disease diagnosis. Car accidents and disease-related deaths are two of the leading causes of death, so these advances would be significant.
Musk countered Zuckerberg with a tweet written in the wee hours of the next morning: “I’ve talked to Mark about this. His understanding of the subject is limited.”
The debate stands.
While I know that Musk’s companies are leaders in the application of artificial intelligence and machine learning and I’m positive that he has in depth, confidential knowledge about these technologies, I think his worries about the consciousness of robots are unfounded – and as unrealistic today as they were one year ago.
In 1955, when the most notable scientists in their field announced from Dartmouth College that the “problem of artificial intelligence is solved,” the solution was where it stands today. We’ve come a long way with significantly more advanced tools and have been able to generate real results from the application of artificial intelligence in the intervening years, but the machine learning component has not developed further. These algorithms have their limitations and cannot achieve the learning that a conscious being can.
For example, deep learning is an effective tool if we are working with an enormous amount of data, but it remains just a statistical method that is unusable if we have a smaller amount of data, and it is certainly not an omnipotent tool.
The consciousness bound to the human brain (or human organism) is not algorithmic based. It seems we don’t store words or the rules that tell us how to manipulate them. We don’t create representations of visual stimuli, store them in a short-term memory buffer, and then transfer the representation into a long-term memory device. We don’t retrieve information, images or words from memory registers. Computers do all of these things (even artificial neural networks), but organisms do not.
There are fascinating new areas of machine learning and I can’t imagine more exciting research than the application of reinforcement learning in time series – and figuring out how the real-world’s rules affect on each other – but I know that we are a million miles from creating a conscious machine – if that end goal is even possible.
If I’m wrong and the robots do rise up in my lifetime, please Miss. AI or Mr. AI (who can read and understand these words), find me and punish me for my own incorrect prediction. It was not made by algorithms.