[LINK] Most consumers hate the idea of AI-generated customer service

[Craig Sanders]

On Mon, Jul 15, 2024 at 08:50:41AM +1000, Tom Worthington wrote:
> Consumers may hate the idea of AI customer service, but will they use it
> anyway when it is disguised, or when it offers a cheap, less error
> prone, readily available service?

Computer says "No"

But they will have to use it when there's no choice, because the corporation
has sacked all the human customer service staff (because they're more
expensive than machines, which are even cheaper than outsourcing to India
or the Philipines. And worse, they sometimes inconveniently try to actually
help the customers they're talking to instead of being the frustrating
obstacle they're required to be....especially when the customer is trying to
unsubscribe from a service they never intended to subscribe to in the first
place).

Corporations don't want messy humanity or empathy getting in the way of
inflicting their policy on customers.

Also, bots can be programmed to favour corporate policy over consumer
protection laws (or employee protection laws), or haven't even been fed those
laws as training data.

And, yes, humans can be obstructionist too - and they're often required to be
by their employers - but not 100% reliably, while computers can be programmed
to obstruct customers forever, without ever getting bored or acquiring a
conscience or considering individual circumstances.

> As an example, how do you know a human, or AI, is processing a form you
> submit online? How much longer will you be willing to wait to speak to a
> human customer service representative?

Customers know, at least intuitively, that regardless of the current
propaganda blitz trying to convince them that this "AI" push is being done for
their benefit, it's no more for their benefit than HR exists for an employee's
benefit.  HR exists to protect the employer from employees and inflict company
policy on them, and the purpose of "AI" "support" is to do the same to
customers.

(HR will be replaced by bots too - they sometimes have too much humanity left
in them)

> If the AI makes fewer errors, do you really want a human?

What a corporation considers to be an "error" and what the customer considers
to be an error can be entirely different and incompatible things.

Especially when the so-called "AI" isn't even remotely similar to an actual
intelligence, but is just really good at regurgitating semi-random strings of
text (or audio & video) in response to prompt.

People wanting support don't want a jumped-up Markov chain repeating corporate
talking points at them, they want help.

craig