I am a writer. That means a lot of my writer friends are worried about losing their jobs. In fact, no one will say: "We fired people because we could replace them with ChatGPT for €20 a month."
But it is already happening.
That worries me too, but I'm not breaking a sweat over it yet.
Because? Because, as of now, ChatGPT is doing a shitty job writing nonfiction. That doesn't stop editors from using it, of course. But readers are starting to notice that something is wrong, complaining that the "how to" story they just read has become a "how not to" story or that an AI-generated article reports that one of the named persons is dead. - when they are still very much alive.
Ups.
I'm not a fan of ChatGPT and other generative AI programs like Google Bard and the recently revealed Meta LLaMA. However, at the moment they are still quite dilapidated.
Tomorrow may be a different story. But in fact, I expect them to be pretty good a year or two from now and great by the end of the decade.
For now though, if you plan to rely on versatile AI to deliver quality work, you're in trouble.
Without a doubt, technology can help. For example, I wouldn't hesitate to send a transcript of a speech, transcribed by another AI program, the excellent Otter.ai, to ChatGPT and tell it to summarize it. Makes sense.
I'm not asking ChatGPT to write anything new. I only ask that he take what is in front of him and make a decent summary of it.
However, what ChatGPT cannot do is give context to what you give it.
That's where people like me come in. In areas where I am an expert, such as Linux, open source software, and cloud computing, I can take that summary and explain where the speaker went wrong or failed to mention that their "big new discovery" had been attempted before by a company. competitor. You had the idea.
Or, for that matter, take Otter.ai. It does a good job, but not as well as an expert human transcriber. So if you need the best transcription possible, say for a legal case, you don't want to depend on it.
Let's look at another example of AI that is not as polished as we would like or fear. The Wall Street Journal recently published an article about Charlie from HomeServe. He is an AI-powered virtual agent used in HomeServe call centers to “answer 11 calls a day, route them to the appropriate departments, handle complaints, and schedule service appointments.”
It does this by using the usual AI tricks, then whispering responses into the agent's ear or displaying their comments on their screens. It can be very useful.
It can also be very boring.
For example, when someone calls to request a repair, they don't want, oh how they don't want, a record spiel.
Of course, it makes sense that Charlie would handle the more mundane tasks like, "Here's the plumber closest to the customer's house." But at this point, Charlie still makes mistakes.
Management seems to like it, but agents… not so much. In particular, the good guys don't want Charlie to guess them. And they really don't want Charlie to rate his performance.
I can't say I blame them. Some people are under the illusion that AI programs are objective and not subject to bias. That's not true. In addition to getting the facts wrong, ChatGPT also responded with sexist, racist, and offensive comments.
We are far from solving these problems.
Right now, everyone is so excited about AI that they can't help but try to use it to solve every problem in sight. So they try it out, and people like it at first because they're new and they save money.
However, over time, your customers realize that what they are getting now is worse than what they were getting before.
Yeah, maybe you cut some costs. Good for you! But you also alienate the people who gave you money. It won't end well.
Used correctly, AI can be useful even at a relatively early stage. However, don't think this is the answer to all your current problems.
Use it to suggest responses to your call center employees. Use it to compliment your writers. In this way, both can take advantage of the AI and avoid the pitfalls that lie in wait for their less suspicious competitors.
Copyright © 2023 IDG Communications, Inc.