Artificial Intelligence (AI) has been a staple of science-fiction for decades, and ‘around the corner’ in technology circles for just as long, but a confluence of recent developments might see it arriving sooner rather than later. The combination of cheaper parallel computation, better algorithms and big data makes AI more likely now than it ever has been before.
Since 2009, over $17 billion has been invested in AI and AI-like technologies, with private investment in the sector expanding by an average of 62% a year for the past four years. Both Facebook and Google have in-house AI teams, and numerous other big names have purchased AI companies in the last year, including Yahoo, Intel, Dropbox, LinkedIn, Pinterest and Twitter. Whether you’re convinced by the concept or not, the giants of our digital era are, and they’re investing heavily.
Commercial AI, however, is unlikely to be as our collective imagination has envisaged it. AI will not produce self-aware, hyper-intelligent beings, but will rather consist of networks of extremely specialised learning computers, controlled by an interface which we will interact with. There might be the appearance of a single system that is capable of everything, but that will ultimately be clever marketing. With the advent of cheaper and ever more accessible cloud processing, such networks are likely to be powered and provided in the Cloud, serving thousands of customers all at once – and getting better at it with every new request.
Perhaps the closest AI will come to its science-fiction equivalent is in what we don’t want it to do: experts predict that one of the main collective aims of AI developers will be to prevent ‘consciousness’ in their AI. The definition of consciousness is a philosophical question, but pragmatically it is undesirable to have parts of our network arguing with each other. Consciousness is counter-productive. In the future, our AI networks will likely be marketed as ‘consciousness-free’.
In fact, rather than being a game-changing cultural shift, AI is likely to be relatively mundane. It will become an underlying system which we take for granted, powering our cars, homes and cities, rarely noticed except for when it goes wrong – much like our Internet connections today.
Whether or not the systems described above constitute true Artificial Intelligence is a matter for semantics, but they are what the term AI will come to represent in the future, and what we will increasingly see becoming part of our daily lives. Data is the currency of the age, and computation the means by which we extract value from it: AI is the innovation that will take that process to the next level.
Image courtesy of Chris Isherwood