The AI revolution – are we losing the human touch?
Recently, DARPA, an agency of the U.S. Department of Defense responsible for the development of emerging technologies, released a video of an engineer bullying a robot. After watching it, I tweeted, ’I wouldn't want to be hockey stick guy the day the robots rise up. They are coming for him first.’ https://tnw.to/d4llC.
Although tweeted in jest, my tweet certainly echoed a strong sub-current of opinion running through the internet right now: where is Artificial Intelligence (AI) heading and should we be worried?
Could AI eventually outperform humans? Is this the point where we become afraid, very afraid?
In a recent Reddit chat, Bill Gates voiced his own fears, saying: “I am in the camp that is concerned about super intelligence. First the machines will do a lot of jobs for us and not be super intelligent. That should be positive if we manage it well. A few decades after that, though, the intelligence could be strong enough to be a concern.” Gates’ sentiment is shared by the likes of Elon Musk and Stephan Hawkins, who have also voiced their concerns over the direction AI is taking.
What risks lie ahead for longer-term artificial super intelligence? The answer is not clear, but I am less pessimistic than Bill Gates and his friends, and tend to side with Rollo Carpenter, creator of Cleverbot who says: “No one knows what will happen if a machine exceeds our own intelligence, so we can't know if we'll be infinitely helped by it, or ignored by it and sidelined, or conceivably destroyed by it. We are a long way from having the computing power or developing the algorithms needed to achieve full artificial intelligence, but it will come.”
One of the biggest challenges for AI scientists to overcome is building Emotional Intelligence (EI) into machines. The emotional complexities that characterise human interaction and decision-making are such that no algorithm can, currently, replicate or master. And we are still a long way from this.
Yes, progress in AI is unstoppable, but as machines are at the mercy of the scientists and engineers who set the development parameters, the ability of AI to take over all human functionality in the long-term, depends entirely on humans. They cannot get ‘smarter’ without input from humans: and then it is a question of values. As Gary Marcus from the New Yorker says: “If machines eventually overtake us, as virtually everyone in the A.I. field believes, the real question is about values: how we instill them in machines, and how we then negotiate with those machines if and when their values are likely to differ greatly from our own.”
But for now, AI doesn’t yet contain the ability to negotiate or interact on a human scale – you only need to look at Apple’s Siri for proof of this, or for that matter, the Robot in the DARPA video because if that Robot had any emotional intelligence programming at all, it would have thrown a left-hook at ‘hockey stick man’! Apple’s Siri has over 30 engineers dedicated to developing its AI, so it can instantly tell you how to find a plumber in Randburg, but it still can’t actually organise a real plumber, negotiate a discount, make sure that the plumber arrives on time, gets the job done nor can it sort out payments between the two parties. The value of Emotional Intelligence is obvious, but AI still has a long way to go. How then do we create a product that will blend the usefulness of AI with a human touch?
This is where, what I call, ‘Assisted Intelligence’ steps in: combining of the efficiency of AI with the ‘intelligence’ of human interaction.
In my day to day life I design and build a mobile application called Hey Jude – a personal assistant on your mobile phone, backed by humans, supported by AI. We use AI to streamline interactions back to a center, where Hey Jude agents make more informed decisions to assist users or respond to a user’s request. For example, the app knows that its client has a meeting in one part of town, followed by another less than an hour later, on the side of town. Hey Jude checks the traffic, distance and projected time to get to the next meeting. If it calculates that its client is unlikely to get there in time, the system alerts a Hey Jude agent who then personally communicates this to everyone involved, and either postpones or cancels the next appointment. In the world of personal assistants we all want to know that our instructions are being acted on accurately and efficiently by a real person – that is why it is called ‘personal’.
In the end, while we all laud the efficiencies and knowledge that comes with AI, we still also rely on human interaction to get the job done properly, or with, well, intelligence.
Hey Jude app is currently in limited release beta. If you are keen to try it out and see how Assistant Intelligence can work for you, head over to www.heyjudeapp.com and sign up to become a tester.