Artificial Intelligence, Real Fear.

Postcards from the Edge #1

I have lain awake for too many nights, wondering who wins in the end. Is it humans, or is it robots?

You see, my work involves behavior change: understanding how and why people embrace new technologies, and this means I am completely obsessed with the frustrating fact that I will not live to see how it all turns out. In fact, assuming the youngest person reading this is 14, and the oldest is a robust 99, then all of our collective consciousnesses will span a paltry 185-year time period at most. When you think of our collective span in the context of the story of life on this planet, that’s a very myopic view. We are all in the cheap seats — none of us will know how this pans out.

So, at my company, we decided to see if we could extrapolate anything at all from the restricted temporal vantage point we are all trapped in. I wanted a way to cantilever out beyond our time to see if we can get a sense of how it all ends up. To do this, we ran a study asking 1,200 people — representative of the US census — a big battery of attitudinal and behavioral questions around technology, robotics, and AI. The sample was large because we wanted to be able to see differences by generation, gender, political and religious beliefs, technology adoption, job function, and even personality traits.

To look into the future without knowing what it holds, we brainstormed a range of applications of AI from the factual (AI chatbots) to the far-fetched (Robot/AI Border Patrol), on a spectrum of mundane (Robot/AI house cleaner), to mischievous (Robot/AI Sex Partner) to the outright macabre (uploading your brain before death so it could live on online).

The results showed quite firmly, that we are not ready for what lies beyond the edge. Within that lack of general readiness, there are also large differences in the level of comfort between cohorts across genders, generations, users of certain products, and more. The first thing to note is that not everyone shares my desire to see how it all turns out. We are quaintly wedded to our place in time.

84% said that they think it is better living now than it would have been a hundred years ago. But only 41% believe living a hundred years from now will be better than now. The rest of us want to hold here. We’ve had quite enough advancement, thank you.

In fact when asked if they were excited or not about AI and robotics a sobering 3/5 were decidedly unexcited, probably because 1/ 3 are worried about robots taking their jobs.

When looking at levels of comfort (in green) and levels of discomfort (in red) across all of the scenarios we proposed to those surveyed, the results look like this:

The first thing to note is the quantity of red. Hence the title: the Dis/comfort index. The red lines represent those who are somewhat, very, or extremely uncomfortable with the scenario in question and the green are those who are somewhat, very, or extremely comfortable with it.

There are only two applications of AI which the majority are comfortable with; house cleaning and package delivery. Concepts like AI lawyers, teachers, etc. are in the middle, and what makes us most uncomfortable — dead last — is the idea of AI/Robot childcare. From an overarching perspective, it would seem that we are quite ready to offload chores (cooking and cleaning), may be open to AI services (lawyers, teachers and financial advice), but are currently firmly closed to the idea of AI care (doctor, nurse, childcare).

This shows, generally, a rather dim view of the future of AI. Perhaps humans prevail after all? However, upon looking more deeply at the data, it turns out we may be more open than we think.

First up, AI has a branding problem. 1 out of 5 who rejected the idea of an AI matchmaker had, in fact, tried online dating services that employ AI-powered algorithms. 45% of those who were deeply uncomfortable with the idea of an AI assistant own devices featuring Alexa, Google Home or Siri. A full 80% of those who rejected the idea of an AI plane with a pilot backup had, in fact, flown commercial. All said, it is abundantly clear that in the world of tech adoption benefits and branding trump fears and qualms every time.

Second, not everyone is fearful, there is a nuance in the neutrals. While the percentage of reds clearly outweigh the greens, it is important to note that between 15% to 25% of people surveyed responded with “neutral” across the board. This group, those who didn’t commit to either side of the comfort range, will likely be the most important segment for the acceptance of AI in the future. As any political campaigner knows, flipping this ambivalent middle often changes the game.

This post is #1 of 5 in a series of Postcards from the Edge based on the findings in our Dis/comfort Index, presented at TED 2019. In the days to come we will publish additional postcards of our findings on various topics.