Let’s pause for a moment, and take linguistic stock.
Technology has many societal impacts, but perhaps one of its most visibly evident is how the tools we use change how we communicate with each other — so, literally, how technology rewrites language itself.
We see this in the willing stockpiling of techie jargon into individual vocabularies (the Androids, the iPhones, the lesser-seen Surface Books); the new word coinages (blockchain), the new and expanding uses for existing words (a ‘message’ is a noun; ‘message me’ an imperative verb asking someone to send you an IM); the creative portmanteaus (glasshole is a particularly good one), the tech-obsessed idioms (I don’t have bandwidth for that right now), and so on and on.
We also get to glimpse the places where technology might be in the process of placing stress on social structures or societal norms. Again, glasshole is a great example of that — a word that effectively contains the full story of Google Glass’ abject consumer failure wrapped up in those two mocking syllables. As I’ve said before, it really does pay to listen carefully.
Well, here’s another interesting phrase that just popped into my sightline as part of a business pitch: “real humans”. The full slogan for the travel-related product it’s attached to reads: “Real humans and smart technology make travel easy and fun again.”
So you can see the writer is contrasting the human brain and machine intelligence in order to claim that, by having both, this 24/7 concierge travel assistant service offers the best of both worlds. Which, in marketing terms, surely checks all the right boxes: they’re saying they’ve got your back.
But at the same time the phrase “real humans” is an oxymoron. All humans are in fact real. So the implication of the phrase is that, in the not too distant future, we the consumer might be being sold interactions with ‘humans’ when in truth we’re chatting to an artificial… (Indeed, in some edge cases, this may have already happened.)
The Turing test that challenges software to display indistinguishably human intelligence to a set of human judges hasn’t yet been passed convincingly — unless you count a 2014 experiment in which a chatbot posing as a 13-year-old Ukrainian boy convinced a third of the judges it was a real boy after five minutes of text chatting.
To most reasoning people that’s far too low a bar to represent anything more than a clickbait headline. Or, perhaps, a commentary on the social engineering of people’s perceptions of children and non-native speakers.
The notion of a chatbot or virtual assistant that can pass as a ‘real human’ for any convincing length of time is not yet credible. (Although humans have shown they can teach AIs to parrot their own prejudices incredibly quickly.) Chatbots and virtual assistant AIs may have got a lot savvier than their Clippy forerunners, but they’re still sitting on a set of definitely discernible rails.
Indeed, our expectations of interactions with any of the current gen crop of ‘smart assistants‘ — at least as a general rule — are set so charitably low that any kind of basic utility, telling us the weather when asked, playing a song we want to hear, is considered a success. When really it’s just twiddling a knob in verbal form. But when it comes to complex last minute travel booking nightmares? Well, we’d really rather talk to a human to fix that.
The goal of technology being indistinguishably human remains a very far distant one indeed. Which means there’s apparently real boastworthy advantage in employing human helpers for certain tasks — especially when other companies might be trying to palm their customers off with subpar interfaces that just end up wasting their time. (On that front, this is a great read on why AI assistant meeting schedulers are too dumb to trust right now — well, unless you’re actively trying to make someone feel they’re not worth your time.)
Coincidentally, just after encountering the promise of “real humans”, an email from an analyst firm lands in my inbox — touting a piece of forthcoming research in which the firm predicts that virtual digital assistants will overtake the world population by 2021.
Should this forecast come to pass, within five years there will be some 8 billion AI assistants out in the wild, mechanically asking us humans if they can help with our myriad, big, small, basic, finicky and all too human tasks.
Such an incoming tsunami of artificial noise might well explain why someone is already spying mileage in employing — and boasting about deploying — “real human” assistance as part of a differentiating business advantage. After all, there’s a reason why the word ‘robotic’ is a pejorative adjective. And will still be for a long time to come.