Five Skills You’ll Need to Collaborate with Artificial Intelligence
July 28, 2017
Ge Reports Canada
These AI experts share key attitudes and behaviors that will help you work effectively with AI.
All of us are already interacting with AI all the time, like when we start making plans with a colleague over email and the software offers to set up a calendar invite.
That’s just a small example of an oncoming tsunami of AI applications that are becoming increasingly integral to how we do our jobs.
Look at GE. Whether it’s designing jet engines, gas turbines, or medical scanners, GE is pairing human operators with deep learning in machines to improve outcomes. GE is deploying these AI-powered solutions on Predix, a secure, Cloud-based platform that allows companies to interact with Digital Twins of their assets on any device.
The ubiquity of AI means we need to get better at cooperating with it. The point is not to compete, but to collaborate.
Masha Krol is an AI experience designer and Gabriel Duford is Senior Vice-President of Development and Technology at Element AI in Montreal. Here are five skills they recommend for getting the most out of your working relationship with AI.
What is your reaction when you call to activate a new credit card or mobile phone and the automated system asks you to choose the options vocally? Most of us still feel silly talking to a machine, especially if it makes us repeat ourselves.
But if we want to work more effectively with a symbiotic artificial intelligence, say Krol and Duford, we’ll need to get comfortable talking to machines.
AI works best if there is a continuous feedback loop between you and the system. Take voice activation: nearly every connected device these days has a conversational interface. When you ask questions to the AI, the system learns and improves.
Voice recognition and natural-language processing are big areas of focus for AI developers and designers. We may feel ridiculous now, but talking to machines is an important part of the future of work, so it’s a good idea to embrace it as it comes.
Symbiotic artificial intelligences are built to amplify human cognitive capacities and to assist in decision making, say Krol and Duford.
As humans, we can usually justify the thinking behind our decisions, unless they are based on gut feeling.
The fact is that in some cases, deep learning models can make better decisions more quickly than experienced humans at the same task.
One challenge is that the current AI technology often can’t explain its thinking process. It has been demonstrated that AI is capable of considering many more variables when making a decision, but we are not always open to accepting those decisions. We must develop the humility to accept that we don’t need to second-guess the AI.
We all have our own unique approaches to executing tasks. These patterns are dependent on our knowledge, our experience, and the way we think, suggest Krol and Duford.
When we work with someone who has a different style, it can sometimes lead to animated discussions. In every case, there are opportunities to learn and improve our decision-making process.
To foster that collaborative learning environment, we have to stay flexible. It’s the same with AI.
People with a great capacity for adaptation will quickly adjust how they approach tasks and decisions to get the most out of working with an AI. It’s easy to get defensive about the good old way, but we can’t afford to skip the advantages offered by AI.
If you wanted to advance science in a particular domain a hundred years ago, you needed to know a lot less than you do today.
The Internet offers a huge knowledge base, but we still have to know what to search for, learn the material, and apply that knowledge to (hopefully) discover something new and important.
A symbiotic artificial intelligence—one that is globally connected and constantly learning—could be designed to provide us with the continually evolving knowledge we need to do our jobs, argue Krol and Duford.
Imagine a full-time coach that possesses all knowledge that currently exists. In order to work well with such an AI, we will need to accept that we don’t have to know everything ourselves. Instead, we’ll let the AI do the heavy-lifting, and build on that base of awareness, sort of like the “Yes, and…” rule of improv.
Let’s think about solving problems in a brand-new way, suggest Krol and Duford: take a completely different approach, redesign a system. These inventive approaches take imagination and creativity, which AI won’t possess (at least not for a very long time).
Creativity can be developed with practice. Projecting into the future, it’s not hard to imagine that the most creative people will be the most effective with symbiotic intelligence.
The most productive thing that AI designers can do is liberate our time. Automating repetitive manual tasks, as well as taking on more advanced assignments—like automatically scheduling equipment maintenance based on real-time data— lets people focus on things that they are good at, like using their creativity to dream even bigger.