New technology is always a hard sell to a change-averse public. But to win support for artificial intelligence is to battle a centuries-old stigma born of celluloid, gossip and the darkest imaginations. In this article, SwiftKey’s Sarah Rowley describes the communication principles needed to allay these fears and show AI as a force for good.
Launching a game changing technology is equal parts exciting and infuriating. Whereas incremental innovations – an improved version of something that already exists – are often lavished with praise; the introduction of a brand-new concept with potential to make a serious, if positive, impact on our lives carries with it a distinct sense of public unease.
This isn’t a new problem. The introduction of steam trains, the vanguard of the industrial revolution, caused a near-on national panic in Britain. The human body, it was said, was not designed to travel at those speeds. According to rumour, passengers could expect to expire in the grisliest fashion—melting was a popular warning.
The telephone was treated with no small amount of suspicion by early users. Some people feared that it would electrocute anyone who touched it (which it didn’t) and that it would unleash an epidemic of gossiping (which it did).
Every new technology can expect to experience a stall in its growth because of poor public perception. But, as these examples prove, it is rarely insurmountable.
Artificial intelligence is unfortunate enough to have an even more punishing adversary than the mainstream media: Hollywood.
As damaging as rumours of electrocuting telephones may have been to initial take-up, create fertile ground for plotlines they did not. (It’s hard to imagine a story about bunch of murderous rotary dials making for a box office hit.) Artificial intelligence on the other hand, has been incredibly lucrative for authors and filmmakers alike for over 200 years. Indeed, ever since Dr Frankenstein’s ill-judged experiment, the appetite for science fiction – particularly the dystopian horror that mankind will one day be the victim of its own genius – has been pervasive.
We’ve seen robot doppelgangers (Metropolis), scheming satnavs (2001: A Space Odyssey), job-stealing public servants (Robocop) and intelligent machines which trap humans in a permanent dream state in order to harvest their bioelectricity (The Matrix).
No wonder people are terrified.
The problem is that the global conversation has got stuck in these Hollywood tropes, a sensationalised bombastic language which fails to recognise the real positive opportunities being made available to us through AI.
We read this month that two robots being tested for Facebook went off script and started talking to each other in their own language. It’s perfect clickbait but by no means an accurate representation of what happened.
For every apocalyptic story, we need a counter message which positions AI not as trying to take over the world or take over your minds, but helping in the real world.
We need to change the conversation. Here’s a starter for ten.
Watch your language
AI still comes across as a remote and inaccessible – almost secretive – concept. We need less combative language. Less elitist language. Fewer acronyms. The tech world can be an insular place which can even unintentionally keep the rest of the world out. Everyone who represents an industry has a responsibility to talk in a measured way.
Humanise the concept What actually is AI? What is it doing?
It’s scientists trying to make computers smarter in order to improve our lives. It’s making healthcare far more efficient. It’s doing incredible things for farmers all over the world. It’s using insects as tools to detect pathogens in animals around them. It’s helping scalability and productivity. It’s everything from predictive language apps to self-driving cars to creating ethical farms. A lot of focus is now going into using AI for environment and climate change research.
If we use tangible examples of human benefit rather than just technological capability; if we explain it in an accessible and democratic way, people are immediately less afraid of it.
How we talk about SwiftKey – the text prediction technology which facilitates faster and easier typing – follows this approach. The conversations we start centre on the benefits AI can bring, and we show our users how much AI can help them—the everyday powered by smarter machines.
In 2014, Swiftkey partnered with Intel to upgrade Professor Hawking’s text-to-speech technology. We built him a personal ‘language model’ which learns from him to ensure it predicts contextually relevant words.
This really enhanced that humanising message—showing people how machine learning can help someone with a disability helps them see it through a different, less alarmist lens.
Address the risks honestly
Ultimately, we must accept AI for what it is. It is a massive scientific and societal development that is happening at a rapid pace. It is another revolution. And, as with anything with this scope and scale, there will be flipsides—the story will never be 100 per cent positive.
It’s incumbent upon companies like Microsoft to be acutely aware of the societal, psychological and scientific risks that AI could bring about. To make sure that the technology we create works to augment human capabilities, rather than diminish them.
By having an ethos, a code of principles, checks and balances, we can go some way towards actually mitigating these negative effects; anticipating and allaying some fears in the process.
You can listen to Sarah discussing this in more detail on the latest Eulogy Behind the Headlines podcast here: http://www.eulogy.co.uk/view/behind-the-headlines-the-challenges-of-innovation-adoption/
Read more about SwiftKey’s work with AI: http://www.hostingadvice.com/blog/swiftkey-keyboard-prediction-software-optimizes-text-based-communications/