With AI being the latest buzzword in many sectors, and so many solutions providers on the market now claiming to use “AI” it’s hard to understand who actually has something unique and original to offer. We’re also being bombarded by hype in the news about AI and robots taking over, such Stephen Hawking’s warning or the recent blown-out-of-proportion story of Facebook shutting down their AI programme because it has simplified the English it had been taught. If you’ve been reading these mainstream headlines around the topic, then it’s easy to believe we’re a lot closer to Terminator than we actually are!
“AI could develop a will of its own. The rise of AI could be the worst or the best thing that has happened for humanity.” Stephen Hawking
What I have learnt this week however is that most AI today is being developed for specific problems. Whether it’s to drive our cars, respond to search queries and questions, sort data or improve personalisation, AI can be a highly valuable tool but it’s a long way away from replicating human intelligence.
We started this week’s session with Dr Janet getting us all up-to-speed on exactly what AI meant (you may think you know but I encourage you to read on.) She explained that a lot of confusion around AI today stems from the fact that the term has so many interpretations. “AI” could technically be used to describe anything from the 20th century definition of “AI enables machines to respond on their own to signals from the world at large” to Artificial general intelligence (AGI) which is “intelligence of a machine that could successfully perform any intellectual task that a human being can”.
This means that when we say AI we’re not being clear enough. We could be referring to anything from a traffic light system that has been taught to sense waiting cars and respond, to Skynet, the fictional artificial general intelligence from the Terminator sci-fi films!
This creates a huge disparity between AI solutions on the market today. Storystream, for example, are working on AI solutions to analyse customer’s behaviours on social media and characterise people into “tribes” rather than demographics, for brands to successfully target with their message. They analyse people’s behaviours, posts and interactions with brands through many features within a photo and look for patterns which can sort customers into particular groups. This is a perfect example of using AI to solve a problem – in this case too much data for a human or team to manage effectively. This is an applied use of AI to solve a problem, which is obviously much more effective than simply advertising “AI” as an added extra on a long list of services that can be offered without clear direction.
I asked her what she thought about the publicity around marketers losing their jobs to AI and robots. Her thoughts were that most jobs and skill sets are constantly evolving, and this area of business is no different. 10 years ago, we didn’t have Instagram, influencers or Facebook Live or but marketers have all adapted their practices in order to use these as effective tools. This will no doubt be the same in the near future, when AI, robots and chat bots becomes an embedded tool for many. In reality it will be lower-paid jobs and developing countries who will lose out on business, as AI helps simplify repetitive and well-defined tasks such as data sorting, researching and processing.
My favourite part of the session was when we began to delve into the impossible, never-ending questions surrounding the definition of human intelligence and the ethics attached to creating, sustaining or ending the life of “artificial” intelligence. Dr Janet got us thinking about creativity as a current limitation for AI (and a benefit of humans for now!) By this we mean while AI can process information and instruction it has been taught and apply them to other situations, it is not able to come up with new ideas or thinking as we do. However the question we discussed was once we truly understand what “creativity” as a skill is, then will we be able to teach it to a robot? Is it therefore a current limitation of our own intelligence instead?
“We have to understand what is human intelligence before we can understand Artificial Intelligence”
You can read her full AI predictions for 2018 here.
This article first appeared on LinkedIn.
If your interested in finding out how AI can help you, we’d love to show you.