In a recent episode of Intelligence Squared, Martha Lane Fox and Mishal Husain, reminisced about the early days of the Internet and, specifically, Lastminute.com, the pioneer travel and leisure website, launched by Lane Fox and her business partner, in 1998.
The journalist Mishal Husain recalled:
“I remember so well that the first two websites that I ever used were Amazon and Last Minute. (…) I remember so well the experience of being able to buy something online, for the very first time”.
To which the entrepreneur replied:
“Well, you say ‘the experience of being able to buy something’. What actually happened is that you would have hit a random bit of the screen. Then, a fax would have gone to another machine, which I probably would have taken out of the fax machine, and faxed into another machine. That would have gone to a supplier that I would, then, have called to say: “Did you get Mishal’s order?’. Buying online didn’t look very much like buying online”.
Hearing this reminded me of a story in the Guardian about how Artificial Intelligence is performing below expectations in several service applications, from transcribing invoices, to replying to e-mails, converting voicemails to text messages, or pretending to be chatbots. As a result, the technology is being routinely complemented – or even replaced – with humans, though those companies going to great extents to disguise when that happens.
The CEO of ReadMe, Gregory Koberger, is quoted in the article as describing such systems as “pseudo-AIs”, and explains that using humans to do the technology’s job “allows you to build something and skip the hard part early on”.
That is, what we are seeing here is customers (and investors) being drawn to a service proposition because of the promise that will be handled by technology; and, then, humans being drafted in, discretely, to fill in for the technology, until it can live up to its own promise.
Isn’t it ironic?