Such a good and sane outcome. Not only am I happy for the affected customers, I’m also going to file this away in my bookmarks for later use.
One of the main concerns I bring up every time one of the managers wants to throw “AI” at something because it’s trendy is “who’s responsible when it just makes something up?”.
I know this is a Canadian ruling instead of a US one, but at least I can point to it and say “probably us”.
This guy was only saved because he saved the screenshots.
This may seem like overkill, but whenever I’m dealing with customer service for a significant financial concern, I always take detailed notes: time, date, name, what was discussed, and screenshots when relevant.
Just a few years ago a large appliance company tried to dick me around on a faulty product that was several thousand dollars.
It took weeks of back and forth, but when I was able to make contact with a layer of management just outside and above their customer service channel with my very long and detailed log of their bullshit, they authorized a check the next day.
Are screenshots even still considered evidence? They should be absolutely trivial to manipulate.
Yes, but submitting a manipulated screenshot in court is a whole different type of fraud and can carry whole different types of penalties than just claiming X happened when it was Y, so very reasonably puts the onus back on the company to look through their logs and try to prove what’s in your screenshots didn’t happen that way.
Good. I’m sure the chatbot will be back up and running soon, but anything that reminds companies there are risks to replacing humans with “AI-enhanced” chatbots is good. Unfortunately, I’m sure the lesson companies are going to take away from this is to include a disclaimer that the chatbot isn’t always correct. Which kind of defeats the whole point of using a chatbot to me. Why would I want to use something to try and solve a problem that you just told me could give me inaccurate information?
Expedia does it right. Just stop stupid customers at the brigade, and connect people with the real need immediately to real people.
Amazon is similar but the real people there are useless as fuck in my country. They’re foreign part timers barely speaking the language if my country… Can’t do anything specific.
Yeah, I bet now we’ll be seeing some real people in chats while they scramble to cover their asses.
There are two disturbing tendencies being demonstrated here:
- Using useless AI to engage and disperse complaining customers. The AI can’t make meaningful solutions to many customer complaints. But companies use it to annoy the customers into giving up, so that they can save the cost of real customer support.
- Either blaming the AI or insisting that it’s right, when it makes a mistake. AI by nature is biased and unpredictable. But that doesn’t stop the companies from saying ‘the computer says so’.
These companies need a few high profile hefty penalties as a motivation to avoid such dirty tricks.
