Chatbots need to make sense, not nonsense

I just had an interesting and lengthy dialogue. After booking flights a while ago, the flight operator cancelled one leg of my flight. Instead of flying to Las Vegas for five nights via Frankfurt, I was now flying to Las Vegas for two nights and then back to my layover city Frankfurt for three nights. WTF?

I call the airline (which took ages).

“Hello, how can I help you?”, says the CSR (Customer Service Rep).

“Well, I booked a flight with you a long time ago, and now you’ve changed my itinerary so that I fly across the world for two nights instead of five”, I say.

“Let me have a look. What’s your booking reference?” says the CSR.

We go through the process.

“Awww”, says the CSR, “it’s a bot”.

“A bot?” I ask.

“Yes”, they say, “a bot. A bot changed your flight because your original flight was cancelled, and so it booked you on the nearest flight.”

“The flight that gives me two nights in Vegas and three nights in Frankfurt, instead of five nights in Vegas and no night in Frankfurt?” I reply.

“Yes”, says the CSR.

“Well, that bot is pretty stupid”, I say.

They agree and we change the flight.

I then sat back and wondered about bots. A bot is an automated view of the world, and it reinforces my discussions of Goldman Sachs and the box-ticking culture.

Box tickers take a view that has no human involvement. It says you are good to go or just say no. They say that a flight that makes sense to the bot, but not to the human, is fine. A human knows that if you book a flight to Las Vegas via Frankfurt that your destination is Las Vegas, and not many nights in Frankfurt. I’ve got nothing against Frankfurt, but just wonder why a bot would do this?

Well, it’s all about GIGO, or Garbage In and Garbage Out.

A bot is only as good as the people that programmed it. We’ve seen this with finance – Goldman Sachs telling most women to get lost, for example – but it is becoming more rife. The latest example is what happened with Meta (Facebook to all those who aren’t with-it). Meta/Facebook launched a new bot the other day which told the world the company exploits people and that Mark Zuckerberg is unhelpful.

OK, there are some bots I love and some that I hate but, either way, bots are also idiots right now. Remember Microsoft's one? The one that became a rascist asshole in just a few hours?

We need true intelligence. It can be artificial, but when a bot says the company it works for exploits people or books a flight that makes no sense, then it’s nonsense.

Give me sense, not nonsense.