The question for this blog;
Are Elon Musk’s statements about Artificial Intelligence (AI) accurate and therefore frightening?
And should we care?
Introduction
I know in my last blog I said I would discuss the debate between Elon Must and Mark Zuckerberg, but on reflection, I could not fully detect whether Mark Zuckerberg has seen the slight change in Elon Musk’s stance and therefore, are they more in accordance over this matter.
So, I decided to blog about what Elon Musk is presently stating and allow you to draw your own conclusions.
Are Elon Musk’s statements about Artificial Intelligence (AI) accurate and therefore frightening?
Well we all know Elon Musk – former co-founder of PayPal (before it was sold to eBay), CEO of Space X and Tesla.
His creations and companies cover so many different technologies, all requiring large amounts of computational power, from designing and operating returnable solid booster rockets to self-driving cars.
So why is he so against Artificial Intelligence, when the companies he runs, actively use so much of it?
In the light of the amount of AI his organisation use, should we care about the things he says?
Is AI more Dangerous than Nuclear Missiles? As he has said, or am I overdramatizing?
Discussion
Elon Musk was asked to speak at the National Governors Association (NGA) held in Rhode Island, USA (2017) where he discussed Renewable Energy Generation Storage, the Tesla factories (current and future) and Artificial Intelligence.
In the discussion, he was asked “How much do you see Artificial Intelligence coming into the workplace?”
Elon Musk’s response was an interesting one because as part of his answer he said that “Until people see robots walking down the street killing people, they don’t know how to react…” (Musk, 2017)
He then goes on to explain that there should be proactive regulatory bodies set up before AI becomes an issue instead of, as has happened in the past, regulatory bodies being set up after the fact. For example, the setting up of driving tests or the Civil Aviation Authority.
Musk also goes on to say that “AI is a fundamental risk to the existence of civilisation”
He justifies this statement by saying that “Robots will be able to do everything, and do it better than us” (Musk, 2017)
Are Elon Musk’s statements about Artificial Intelligence (AI) accurate and therefore frightening? So, we can infer that Elon Musk is very worried about AI, but what AI is he worried about? Narrow AI, General AI or Artificial Super Intelligence?
Elon Musk interviewed by Jonathon Nalon in an on-stage interview at the SXSW Convention held in Stockholm (October 2018), made the same statements, so we can assume that Musk feels very strongly about this, however in this on-stage interview, he goes on to define the category of AI that is concerning him, that being Artificial Super-Intelligence.
So instead of saying that AI (generally) is dangerous, Musk has defined his argument and is advocating for the proactive regulation of Artificial Super-Intelligence.
Most of us have seen on the news or heard reports that AI is being tested in banks to trade in stocks and shares, in medical facilities to detect cancer in X-rays, and have seen that we can now unlock our phones or front doors just by looking at them.
But is this true Robotic AI?
These technologies do not use robots in any stage of their process, therefore cannot be defined as Robotic AI!
So, what is this AI that we are hearing so much about?
Well it is categorised as Narrow Artificial Intelligence, and this is the learning a machine (in this case a computer) can do to process sounds and patterns quicker and more efficiently than humans are able to do.
Hence, we can use this Narrow AI in face recognition or identifying cancerous cells in scans with higher efficacy than human doctors can.
So, let us ask the question again, but in a different way….
Should we, as Elon Musk is saying, be concerned about Artificial Super-Intelligence?
Well what is super-Intelligence?
In short, and I will be looking at what Artificial Super-Intelligence is in more detail in my next blog, it is a computer or a machine (robot) that can do everything we can as humans.
For example – It can walk and talk, switch subjects seamlessly mid-conversation, and read the social cues that people as individuals give. But it can do these things faster and better than we can.
In other words, potentially be 100% better at everything that we as humans can do, but with less of our faults and frailties.
Stronger, faster, more intelligent, and with the field of prosthetics and robotics advancing as it is – potentially indistinguishable from human beings……………Frightened yet?
This is what Elon Musk is concerned about, that we could have unregulated robots that look and acts as a human being does.
Should we care?
My personal option is that- well yes to a point, we should care. Why?
I do not think we should stop or impede the research into Artificial Super-Intelligence, because it will be this that enables us to remain as explorers and travel further into space, help us make medical and scientific breakthroughs.
But should we give Artificial Super-Intelligence the ability to fire nuclear missiles, well I think that we all might agree that may not be the best thing.
But what is going to stop an Artificial Super-Intelligence robot from going mad and killing people, we this is where the regulation that Elon Musk is talking about comes in.
So, Is Elon Musk right in that there should be some sort of proactive regulatory body set up?
Personally, I agree.
Why? So we don’t have killer robots walking down the street, but what is going to stop the military from connecting the defence systems to an Artificial Super-Intelligence computer
Will Super-intelligence be a civilisation-threatening move?
Well, it could be…. but as no one knows what superintelligence looks like yet, no one can theorise accurately or definitively.
As it stands presently, we still have a lot of work to do combining Super-Intelligence in the head of a two-legged robot, if it is to be contained in the head at all.
Two-legged robots that can walk and run are getting better, but they still have some way to go to be able to properly mimic a human being.
However, a final concept for you to consider….
‘Deep Mind’, part of Google, now has a machine that can play any Atari game from the 1980s and beat them, without being told what to do.
Even though that is a set of games, and ‘Deep Mind’ is housed in a massive set of data centres, can this be thought of as Artificial Super-Intelligence?
It certainly seems a step towards this reality.
Please feel free to share and post your views on Elon Musk’s statements and your thoughts on AI…..
Bibliography
Anon., n.d. BBCTechnology Page. [Online]
Available at: https://www.bbc.co.uk/news/technology-49508091
[Accessed 15 January 2020].
Clifford, C., 2017. [Online]
Available at: https://www.cnbc.com/2017/12/18/9-mind-blowing-things-elon-musk-said-about-robots-and-ai-in-2017.html
[Accessed 20 January 2020].
Musk, E., 2017. Elon Must at the NGA Conference [Interview] (15 July 2017).
Musk, E., 2018. SXSW Conference [Interview] (2 October 2018).
Musk, E., 2019. Elon Musk on Artificial Inteligance [Interview] (19 October 2019).
Next Blog
What is Artificial Super-Intelligence? (And could it start another World War?)
