Just before the summer we attended Humanising AI: an international event discussing the intersection of human computer interaction and artificial intelligence. It was clear that AI and the effects it will have on our professions and futures is preying on many people’s minds – just as the introduction of steam and other forms of automation played on the minds of workers in the past. But is AI any different? Here are four things we learned from the event.
1) We need to tame AI, but will we be able to once we set it free?
Humans like to delegate. We have done so for millennia, so it was a natural step for us to delegate to computers. And it will certainly be an interesting journey to see just how much we delegate to machines as AI becomes more and more present in our day-to-day lives. But no matter how much responsibility we onboard to machines, we will still need to feel that we are the masters of them.
We want to be able to turn off technology and “take back” control – the self-driving car being the classic example. There are certain situations where we want to drive a car ourselves, in the same way that pilots tend to land planes manually even though the autopilot can handle it.
To do this with AI automated tasks, we have to be able to tame AI and build in stop functions. We also have to make sure we don’t forget how to do the very things AI is doing for us once we make the decision to take back control. It may sound unrealistic to think that we will forget how to drive a car, but how many of us know how to hunt and provide food for our families?
2) Somebody is probably looking at ways for AI to do your job
We all know the impact AI will have on drivers’ jobs. Virtually everyday there is a news story about self-driving cars and trucks, but most of us are less well informed about other professions. In the finance industry – where traditionally many of the brightest minds seek employment straight out of university – many technology and operational positions are now at risk. Many trader positions, for instance, are being replaced with computers, but on the flipside of that there is an opportunist market for traders who are betting against the trading algorithms.
For us at Daresay, it was really interesting to hear about the AI advancements being made in design – since it’s a lot of what we do at here. A good example of AI outperforming humans is in predicting certain behaviours and using this to create more effective designs. This includes identifying the most effective positions of website calls to action, or designing games graphics that provide the biggest challenge for gamers. AI can predict human interactions with computers and identify patterns far quicker than we ever could do.
And that’s just web design and gaming. The uses for AI are being looked at from every angle. In fact, in a room full of AI experts and enthusiasts, the only thing we could agree on was that plumbers’ jobs were pretty safe, for now.
3) AI creates jobs, but what kind of jobs?
AI isn’t just replacing jobs (or assisting people at work – depending on your point of view), it’s also creating them. For one thing, there is a lot of coding to be done. But we learned that people are often needed to cover for machines when services are promised but coding can’t be done quickly enough to deliver them. One example given was from an online matchmaking site for service providers and people looking for household services. For a long period of time it was humans that were matching providers with service requests rather than machines because the code hadn’t been created to deliver the service. That same company also had a lot of call center operators. It’s become apparent that people want to speak to other people if they have a problem or a complaint. Interestingly, there are instances when people want to speak to a machine rather than a person. This is typically when they have done something they feel a little ashamed about, such as missing a monthly credit payment – a machine will not judge them, unlike a person.
4) We can’t over rely on AI
There are things AI is good for and things it’s not – and we should never forget that. Accordingly, we should use AI to automate what is simple for people, such as driving a car and put our own skills to solving harder problems. AI doesn’t care about aesthetics for instance, and how an Alexa fits in with the decor of your home – that’s the designer’s job. It’s AI on the inside, human on the outside. Of course AI could be used to design a product, but would it take into account the exceptions or would everything be designed for the masses? The actual actions of one person can tell you as much as the predictive actions of tens of thousands of people, which is why it’s always going to be important to talk to people when it comes to design. By studying individuals actions when sat in self-driving cars, for example, researchers found that people were more attentive to what’s going on when reading or watching TV as opposed to just sitting and “not” driving – which had a tendency to send them to sleep. For this reason further research is being done on the best ways of keeping people alert in self-driving cars.
Much more than this was discussed during the event, including the inevitable introduction of truck platooning and how AI is being used to track “the bad guys” online by constantly scanning new data on the web. One interesting thought to round off with was a conversation on the comparison between AI and the web.
In the early days of the web there were a lot of discussions about making it machine readable as opposed to human readable – because that was where the supposed true value laid. But the web is an amazing social environment, connecting people everywhere. Could this happen with AI? Could we see a dominant use that is contrary to what many of the early thought leaders believed? Stay tuned, we will certainly find out soon enough.