BLOG
blog-banner

Chatbot Confusions Will Blow Back On The Unwary

5th September 2017 No Comments

VoiceSage’s Matt Weil argues that misleading AI coverage potentially harms our appreciation of what this tech can genuinely do for brands

Earlier this month, you may have read a story that gripped the world’s media about a sinister-sounding piece of AI advancement.

We were told that Facebook had had to ‘shut down’ an experiment that, SF horror film cliché, was about to threaten the world.

‘Robot intelligence is dangerous’: Expert’s warning after Facebook AI develop their own language’ and so on.

The claim was that researchers at Facebook had noted the uncanny activities of ‘Alice’ and ‘Bob’ – two ‘bots’ that had created their own secret language that the boffins couldn’t themselves understand.

“When the scientists returned, they found that the AI software had begun to deviate from normal speech and were using a brand new language created without any input from their human supervisors,” was one version of the story – with the reporter quoting an ‘expert’ to the effect that anyone who doesn’t think this is not dangerous “has got their head in the sand”.

However, all the fuss stemmed from an innocuous trial of chatbots in the context of negotiation. Facebook AI coders had been experimenting with bots that negotiated with each other as part of an effort to understand how linguistics played a role in such discussions.

The bots were simply programmed to experiment with language in order to see how that affected their dominance in the negotiation, and they worked out some clever shortcuts to do that.

As Gizmodo says, “In their attempts to learn from each other, the bots thus began chatting back and forth in a derived shorthand.

“But while it might look creepy, that’s all it was.”

No scarily smart AI we couldn’t understand.

Use chatbots wrongly, and the public will take a dim view of you

And of course, too many people don’t understand AI.

For the customer contact world specifically – they are misunderstanding what chatbots (the most immediate and actually eminently practical application of AI ideas to our space) can and can’t do.

We are over-hyping chatbots and expecting too much of them. Early usage of chatbots has in many cases disappointed brands as a result, as they are pushing them to replace humans instead of augmenting them.

Customers are very good at spotting over-eager AIs a mile off, and that spells ‘F A K E’.

If a bot starts to try and chat about the weather when you just want to find out if your delivery will come in on time, the experience becomes a negative one.

Brands pushing too hard to automating far too much of the customer outreach process – seeking cost-savings and efficiency – will face a backlash.

So just as you need to exercise scepticism about these ‘Killer AIs’ stories, you also need to exercise some judicious thinking about AI in customer service.

Don’t force chatbots on to your customers. They are great, but be realistic; chatbots will work as an addition and a support for the contact centre.

Use AI/chatbots to automate simple, repetitive tasks the user needs to do, or as a fallback. Don’t force the user down a path they don’t feel comfortable with – it never works.

Drop me a line if you’d like to hear more about what we’re doing in this promising area of technology and I’d be delighted to explain more.

And in a language we can both understand!

Matt

Matt Weil is Head of Product at VoiceSage