The most mind-bending technologies often outpace what we think of as possible. For example, miniature brains scientists have grown from cultured tissue and used to direct wheeled robots. The scientists hope to discover how human memories are created and then how they may be corrupted by degenerative neuro disease. The same was at once true of artificial intelligence but now AI is not only feasible, it's in your smartphone.
Researchers are experimenting with AI to help deal with serious illnesses (Photo: Author's own)
We need a mature debate about AI implications
AI fears seem to come in two flavors: intelligent robots may take our jobs or they may take our lives. But as we move from the fantasy world of the big screen to the solid sidewalks of everyday life, AI machines have not turned into killer robots. Instead, AI algorithms give fashion tips, send people on romantic dates, suggest music and anticipate other digital desires.
Of course, history has shown that technology disruption leads to both job losses as well as new industries with new jobs. While we can expect the same to be true of AI, we should think about the implications on workers in the short term and prioritize technologies that enhance people's abilities rather than simply replacing them.
AI is already having profound impacts, such as the work being done by Oxford's John Radcliffe Hospital and Coventry University - both in the UK. They are assessing how AI can help control tremors in patients suffering from Parkinson's disease. The idea is to implant an electrical activity monitor into a patient's brain and feed data into an AI device. The researchers believe the AI will predict tremors up to 20 seconds before they occur, providing the system with enough time to activate a brain stimulator and fire electrical signals to nullify the tremors.
Health applications are great examples of why I believe AI will be a key part in a utopian (not dystopian) technology future. Yet, AI today is in embryonic form, with algorithms primarily running in research labs and vast 'big iron' data centers. The challenge for the tech industry now is to push this learning and intelligence into all layers of the system.
Why do we need artificial intelligence everywhere?
Google's Federated Learning is a great example why a distributed AI computing model is best. It uses machine learning to gather data on smartphone apps, using the information to improve the app experience. This could be done in the cloud but by deploying 'AI at the edge' the updates happen immediately and all private data remains on device. Later, the information is aggregated on Google's servers anonymously and used to refine the app for everyone. Win, win.
The personalization of AI will allow the immediate application of learning, enhancing security. It also potentially removes a huge additional data load off the network. Many sectors looking to adopt AI will also force change. AI in cars, as well as industrial and medical robots, all rely on the ability to make instant offline decisions. An autonomous car must be able to think, decide and act, even if it loses connectivity.
Intelligent machines need intelligent design
As well as functionality, we need to enhance the efficiency of artificially intelligent devices. For example, AlphaGo AI computer's victory against the reigning Japanese Go game champion in a four-hour match was compelling. But AlphaGo consumed 500 kilowatts of energy to tip the balance against Lee Sidol. That was the equivalent to AlphaGo consuming 3,000 hamburgers' worth of calories against less than one hamburger by Sidol's organic brain.
Intelligent and secure
A holistic AI system must be secure to win the public's trust (Photo: Author's own)
AI will attract hackers, so we need a common language enabling any trusted hardware to hook into a software security layer, creating a bond of trust. This must work as part of a system-wide solution like the mobile world where network operators police every connecting device. The approach for AI, and more broadly theInternet of Things, won't be identical though as the system must deal with many more devices.
In the years ahead, it's important that we begin the dialogue that helps to capitalize on the potential upsides to AI while mitigating for the short term human impacts on the workplace. As Dale Carnegie said: "If you want to conquer fear, do not sit home and think about it. Go out and get busy."
Suggest a correction