Fri, 07 Sep 2018 10:18:56 GMT
Hal 9000, Rosie the Robot, Skynet and that cute robot from Short Circuit - artificial intelligence has been a topic that has gripped the imagination of people for the past 60+ years. As computers have gotten smaller, faster and cheaper, AI has been able to move from the world of science fiction to science fact. The Watsons, Hive Minds and TensorFlows have flung AI into the mainstream as a viable technology that can impact everything and anything. As such, should we be welcoming our new AI overloads?
Generally speaking, you can break AI research into two main categories - specialised and generalised. Generalised intelligence is the ability to create an artificial intelligence that can solve multiple types of problems; these are your HALS, Skynets and Rosies. Specialised intelligence is the ability to solve domain specific problems - i.e. detecting dogs in photos or recognising your voice. From a technology perspective, we have done a great job at pushing forward specialised intelligence, but not so great a job at generalised intelligence (it’s a much harder problem), as such any practical application of AI must be rooted in specialised intelligence.
Looking at specific problems AI can solve gives us a filter we can use to determine how AI can impact our business, creative and products. This filter is effectively the question: “For a given problem – can we use AI to augment our workforce, business or products to make the user experience better?”
Using that filter, let’s look at three top examples of how companies are using AI to augment their consumers and employees, driving the way forward for AI.
Baidu had a problem. Ad revenues were dropping, and trafficking teams were not able to keep up the demands of their networks. Across the company, Baidu had been moving to an ‘AI first company’, and this seemed like another place where they could apply AI to solve workflow problems.
Working with their digital advertising team, an AI system was implemented that augmented their trafficking and display process. By using an AI system to make real-time decisions and support their core advertising team they were able to grow net ad revenue by 24% from the quarter previous and provide more relevant advertising to their end users. This application of AI – while very business focused - ended up augmenting their workforce to drive efficiencies and free up their team to be more strategic and provide more relevant content to their end consumers.
Flagler Hospital wanted to understand how it could use technology and AI to make its operations - and the ultimate ability to serve patients - perform more effectively. With this objective, they looked at their approach to servicing patients with high-cost illnesses like pneumonia and sepsis. Using AI, they were able to analyse the core pathways doctors were following, and make suggestions on how they can be more efficient. After implementing the changes recommended by the AI-driven system, they ended up seeing a net reduction of hospital stay for these patients of two days, $1,000+ inpatient savings and better operational yield. By using AI to augment the way they process data and results they were able to drive an improved patient experience. While not directly implanting AI in a real-time sense they were able to use it to improve their business intelligence and ultimately drive a far better care experience.
Google (Android P)
As with all modern devices, Google wanted to understand how they could optimise their battery life of Android phones. This was a perfect problem that AI could solve as it could understand user behaviours and optimise the phone to that behaviour. With the release of Android P, Google deployed AI to do precisely this. By understanding what apps a user uses and when, it could optimise how much energy is given to which software. This results in a more efficient power consumption policy and thus longer battery life. Now any device using Android P with adaptive battery enabled has been able to have a ~30% reduction in wake locks (the thing that drains power). This is a consumer-focused application of AI that has augmented the way a user uses their phone to improve battery life.
These three examples are just the tip of the iceberg on how companies are using AI to augment their business, workflows and products to drive better experiences and efficiencies. As we move forward into the wild west of artificial intelligence, we need to continue to apply this filter so we can avoid the pitfalls of hype and deliver better experiences and products to our consumers.
David Justus is executive head of technology at AKQA NY