Invasion of the robo-bodgers
- Summary:
- Every advance in user experience unleashes a wave of experimentation. Today's hotbed of crass innovation is the software bot - the robo-bodgers have landed.
So it is with every wave of technology that opens up new frontiers in user experience. In the early days, it's all about exploring the novel capabilities that designers can build into the medium. It's only later that those early experiments are cast aside, to focus instead on how best to deliver the most useful outcomes. Nowadays, savvy website operators know that visitors don't want to stop to admire their designs — they just want to get stuff done and be gone.
Today's hotbed of crass innovation is the software bot. There's been a sudden explosion of interest in showing off what intelligent software robots can do — much less in whether they actually help anyone achieve anything useful.
All the excitement about chatbots at Facebook's recent F8 developer conference was accompanied by a tsunami of articles about how buggy and unproductive these early examples have proven to be. One writer found it took two and a half hours to complete a simple online shopping task using a bot interface. No wonder my colleague Jon Reed is asking should the enterprise care?
Bodged software bots
The sad news I have to bring you today is that this is going to carry on for quite a few years to come — and it will get a whole lot worse before it ever begins to show a glimmer of productive light on the horizon.
While there's no doubt that AI bots will ultimately become an important — if not the dominant — component of how we interact with computers (which basically means everything we interact with as our lives become increasingly digitized), nevertheless that will not happen until we have become heartily sick of all the ways in which they've failed to perform any useful function at all on the tortuous journey to that promised land. Prepare to be underwhelmed by an invading army of bodged software bots.
Even now, eager developers are laboring away with the Facebook Messenger Platform, launched in beta just two weeks ago, to build their own bots. Many more are exploring the potential of the Microsoft Bot Framework, introduced in preview at last month's Build conference.
Microsoft CEO Satya Nadella told developers at Build that "human language is the new UI layer," part of a concept the company is calling Conversation as a Platform, in which:
Bots are like new applications, and digital assistants are meta apps or like the new browsers. And intelligence is infused into all of your interactions. That’s the rich platform that we have.
And in a throwaway remark that revealed just how crazy all this bot-infused and digitally-assisted experimentation is going to become, he added:
We don’t even stop there. We want to add richness to these bots. So it’s not just text bots. We want to have animation, we want to have videos and we also want to have holograms.
Conversational misdirection
Along the way, these conversational robots will have to learn to deal with all the misdirection humans will enjoy throwing at them until the novelty wears off. As the BBC's technology correspondent Rory Cellan-Jones wrote after trying out x.ai's robotic meetings organizer:
I am sure bots ... have a bright future. But first we and they have to learn how to get along together. That may mean that the humans have to stop teasing, and the bots have to stop pretending they are more than just clever pieces of software.
The key problem here is the main difference between humans and machines. Humans are good at understanding context (or more precisely in this specific case, conversational implicature). Machines are good at providing answers, provided they have already been given the context. They may have reached the point of being able to use natural language, but free-flowing conversation is quite another thing.
So the next few years of bot experimentation are going to be all about discovering where the boundaries lie between utility and time-wasting (along with a good few demonstrations of how badly it can all go wrong). Just like the early days of web design, we face years of trial and error, until developers and users alike finally learn to distinguish between a useful bot and one that's just so much AI candy.