There is a lot of discussion in academic and tech circles about AI. Famous brainboxes like Elon Musk and the late Steven Hawking have been warning about the implications of building truly reasoning machines. We already effectively have machines making critical "decisions" in many areas of life e,g, transport, energy, consumer electronics and so forth. As of now there is no such thing as an "intelligent" machine. There are systems that exhibit certain characteristics we may think of as intelligent but they are currently based on "machine learning".
Google Home, Amazon Alexa, Apple Siri, Microsoft Cortana, IBM Watson etc. are all examples of so-called intelligent assistants where a lot of effort has been put into trying to get them to respond to general enquiries.
However impressive these systems are, they currently all face severe imitations outside responding to predictable questions within certain, quite constrained, parameters. i.e. they all fail the Turing test which is the holy grail of AI. The Turing test imagines a scenario where one can chat to a machine and not be able to ascertain if it is a person or not.
Recently I was involved with putting together a proposal for a 3D "intelligent" Avatar to replace the triage nurse in an ER. We got quite far down the road with the speech recognition side, thanks to some software from IBM Watson, but trying to conceptualise how to initially diagnose someone and the relative seriousness of their injury required us to imagine all sorts of sensors, qualified intelligent responses to questions, awareness of the appearance of the person, assessment of pain level etc etc. Very interesting project but it stopped at the conceptual stage.
I am actually relatively optimistic about AI and its advantages in health, energy, transport - I think the benefits outweigh the risks, even with self-driving cars and aircraft.
Much AI will remain invisible e.g. improving the pictures your phone camera takes, reducing energy consumption, improving transport efficiency and fuel consumption, more efficient mineral extraction, agriculture and so forth. Some will be more obvious and fun or helpful e.g. enabling your phone to be a medical diagnosis or medical condition monitoring tool, automatically getting your drone to follow you and take selfies, better computer games, smart gym machines etc.
I also think that the actual implementation of AI, that is actually visible to most people, will be as "intelligent assistants" at work and in the home. In other words, I do not believe everyone will have to become a technologist or coder. In fact, basic "coding" is likely one of the first jobs that AI could replace!
Google Home and Alexa are only scratching the surface here. e.g. Lawyers are already using AI software to initially "read" long documents to flag where the human lawyers should focus more attention. I see this as the next step in AI evolution e.g. engineers will use AI based software to suggest solutions to complex design issues for consideration by human intervention, or to suggest to farmers the best time, place to plant and harvest crops or how best to reduce pesticide usage. I am just free-wheeing here but you get the idea. I don’t think any of these jobs will go.
Also AI offers the best opportunity in history to revitalise manufacturing because it provides the tools to help distribute design and manufacture where it is needed and/or where people have bright ideas. So, expect to see an explosion in distributed custom product and device
manufacture.
The arts too have never been so accessible; a talented child with a guitar can effectively have their own recording studio (with virtual sound engineer) for less than the price of an X-Box. You can make pretty good movies with a Digital SLR and edit them on an iPAD. You can learn almost anything for almost nothing, thanks to e-learning - languages, plumbing, cyber-security you name it. These things are already happening and barely a month goes by without a conference or event about e.g AI in the financial sector, health sector, transport sector, internet of things, agriculture, food production, 3D printing etc etc.
I will accept that despite my optimism, there are some real potential dangers (i.e. nightmare scenarios) with AI - especially in the area of weapons of war and terrorism. We should also be very clear minded about what we are letting loose in the world and a la Isaac Asimov we should put in place safeguards where possible. The good news is that we are a long way from Skynet taking control and humans becoming extinct, but now is definitely the time to start imagining and working on the best way to use these technologies for the betterment of ourselves and our environment.
I was specifically asked recently about self driving cars - buses are all very well and soon every major city will indeed have electric buses. I like buses and they are very useful in the city where they are frequent and chances are there are ones that are going in your direction. But where I live, the buses go nowhere I want to go. I believe self-driving ride sharing cars are the future. Imagine a future Uber or Lyft where when you call up a ride and it will, by default, be a ride share unless you have a specific requirement. For example, you and 6 friends are going to dinner or you need a large van to take some furniture to your apartment. This will lead, for the first time since cars were invented, to fewer cars on the road. Virtually all cars spend most of their time underutilised. And how many times do you see cars with one person in them? AI will have a big role here to since as well as the self -driving the system will "learn" that fpr example, on Wednesdays at a particular time, a lot of people from that particular neighbourhood go to the dentist, doctor, gym, supermarket, pub and so on. The move away from private cars will happen first in the cities (as is already happening) and slower in rural areas. In fact ride sharing or car sharing is already happening especially in cities.
Of course, I could just be overly optimistic, but in my experience human ingenuity and goodwill are powerful weapons. Things, despite apocalyptic news media reports, are getting better for most people on the planet.
Dr Graham Dodgson
Comments