Get Strategic insights into your business!
We develop and deploy solutions focused on your business needs and objectives
Eleanor Brown, Director of Business Change at Aelm
Artificial Intelligence began to prick public consciousness in the 1950s. Arthur Samuel built checkers, his seminal machine learning program, and Daniel Bobrow launched STUDENT, the natural language processing program. Such breakthroughs led to Marvin Minsky’s prediction that full AI – machines successfully executing intellectual tasks – would be in play by the 1970s.
But instead of marking AI’s coming-of-age, the 1970s were actually the cold years. Limited investment and lowly computation power couldn’t match early promise, and the movement flatlined. A small resurgence came in the 1980s but, for similar reasons, AI and the mainstream remained virtually incompatible.
Before we get drunk on the what ifs of tomorrow, let’s back up a little and soberly consider What is AI anyway?
Enter the 1990s. With powerful computers and falling costs, ubiquitous data and advances in statistical analysis, the innovation bar lowered and folk started to dream again. Fast forward to 2017 and AI as a market is set to double in value to $120 million.
Today we’re told AI will revolutionise our lives. From manufacturing to medicine, banking to building, they say it’ll touch every industry; it’ll steal every job.
Robots and spaceships, autonomous this and self-driving that, it’s equally fun and daunting to imagine a world built entirely on AI. But before we get drunk on the what ifs of tomorrow, let’s back up a little and soberly consider What is AI anyway?
There’s three key components, or steps, that comprise AI: perception, cognition and action.
The easy explanation is that Artificial Intelligence is intelligence exhibited by machines. What’s crucial in the context of AI, however, is that intelligence subdivides into three component parts. It’s rare that all three parts function in harmony, and that’s when we get Hollywood AI: self-determining computers, robotic-best-friends, and KITT from Knight Rider.
That’s the extreme, however. There’s three key components, or steps, that comprise AI: perception, cognition and action. For example, a machine must collect data about its environment (perception); it must process that raw data to decide on an appropriate course of action (cognition); and then it must act on it to affect the environment in some way (action).
When much of the media talks AI, they leapfrog straight to the Hollywood version, think HAL from 2001: A Space Odyssey, where all three steps work as one. And sure, some companies are making significant progress in units like that.
In actual fact, we’re still some way off the widespread deployment of these AI units; much further than Wired would have you believe. Today, AI is best defined not as the sum of its three parts, but as the parts themselves: perception, cognition, action.
And units based on one or even two of AI’s three steps are within much easier reach for companies today. Applications based on either perception or cognition or action are available and ready to plug into businesses that don’t have the cash reserves of a NASA, Tesla, Google or Amazon.
The most common AI systems are algorithm based. Recommendations given by Amazon, YouTube, Netflix and Spotify are AI. So are banks’ behavioural fraud detection systems; so is Siri, and so are the computer-controlled players on games like Call of Duty.
We see and use these AIs already, almost without knowing. And they’re getting better and bigger and more crucial by the day.
The next articles in our AI series will focus on each deployment in turn, exploring how perception, cognition and action are being used, and how businesses can use artificial intelligence now to boost the bottom line.
Published on KNect365 on April 19th