Is the New AI Chatbot the End of the World as We Know It?

Mathematician, computer scientist and famed code-breaker Alan Turing said that if you had a conversation with a computer and couldn’t distinguish what it said from what a human would say, then the computer must be intelligent and in some sense self-aware. The other day I ran a Turing test on ChatGPT, a chatbot recently released by Open AI.

It flunked. But it’s still a student, and it shows promise.

ChatGPT, its makers tell us, is still in beta form. Like a million other new users, I’ve been teaching it (tuition-free) so its answers will improve. It’s pretty easy to run a tutorial: once you’ve created an account, you’re invited to ask a question or give a command. Then you watch the reply, popping up on the screen at the speed of a fast and very accurate typist. Here’s an example:

Me: Write about David Eby in the style of The Tyee.

ChatGPT: David Eby, the BC NDP's attorney general and Minister of Public Safety, is a well-known and respected figure in the province's political landscape. Known for his passionate advocacy for social and environmental issues, Eby has made headlines for his efforts to tackle issues such as money laundering and housing affordability.

Born and raised in Toronto, Eby moved to B.C. in 2003 and quickly became involved in politics, first as a member of the BC Green Party and later as a candidate for the BC NDP. In 2013, he was elected as the MLA for Vancouver-Point Grey, and has since been re-elected twice.

In his role as attorney general, Eby has made headlines for his efforts to combat money laundering in the province's real estate and gambling sectors. He has also introduced new legislation to increase housing affordability and protect renters from unfair evictions.

In addition to his work as a government minister, Eby is also an active member of his community, volunteering with organizations such as the Greater Vancouver Food Bank Society and the Vancouver East Cultural Centre. He is known for his dedication to public service and his commitment to making B.C. a better place for all its residents.

ChatGPT’s description ends here.

ChatGPT’s knowledge base ends in 2021, so it’s not up on recent developments in Eby’s career. But no Tyee editor would agree that it’s “in the style of The Tyee.” For one thing, The Tyee would include sources for both praise and criticism of our new premier, expressed in attributed quotations.

Still, it’s fairly good boilerplate, the kind of text you’d see in plenty of politicians’ news releases.

The Luddites’ response

Early responses to ChatGPT have been largely Luddite: critics have warned that its arrival means the end of high school English, the demise of the college essay and so on. But remember that the Luddites were highly skilled weavers who commanded high prices for their products; they could see that newfangled mechanized looms would produce cheap fabrics that would push good weavers out of the market. ChatGPT, with sufficient tweaks, could do just that to educators and other knowledge workers.

Having spent 40 years trying to teach my students how to write, I have mixed feelings about this prospect. But it wouldn’t be the first time that a technological advancement has resulted in the atrophy of a human mental skill.

Writing arguably reduced our ability to memorize — and to speak with memorable and persuasive coherence. Some “preliterate” peoples expressed their cultural values in huge structures like Stonehenge and Poverty Point — structures requiring detailed engineering knowledge as well as communication skills, passing the “how” from generation to generation along with the “why.” The Incas were astounded when they saw how the new Spanish arrivals could read and write; but the Spanish just took for granted the Incas’ architectural and technical achievements.

Going the way of cuneiform

Similarly, the art of penmanship is going the way of cuneiform. Kids are more used to keyboards than pens or pencils; I can barely stand to watch younger adults clutch a pen as they struggle to write cursive. The same is likely true of mental arithmetic in the age of the calculator app.

Literacy has, um, literally left us all with new ways of learning. The modern classroom is a reflection of those. Students must take notes (or be given handouts) because they won’t be able to remember much of what the teacher said. They dread standing up and speaking in front of the class because they know they’re terrible speakers (their teachers, alas, are rarely much better). Then they have to return whatever information they’ve rented in the form of essays, quizzes and exams before lurching off to another year of the same.

ChatGPT is nowhere near ready to replace modern education. It’s been exposed to millions of words (and is now being taught by millions of unpaid tutors like me), but it doesn’t even connect to Google. Like any impressionable kid, it can be deceived by misleading information and plain lies. I know I shouldn’t take it personally, but I’m still sore that ChatGPT emphatically denied that I’m a science fiction writer, while attributing two books to me that I never wrote.

A trusted companion and oracle

Still, I can see some very useful potential in future versions of ChatGPT and its inevitable competitors. Recall how we all migrated happily from landlines to cellphones to smartphones, which we now consult like pocket oracles. Similarly, we might find a really smart chatbot a boon companion.

But to get the most out of such a chatbot, we’d first need to be the subject of its attention. We’d have to go to school for years, developing skills in reading, writing, speaking and listening. Once we had a body of such work, we could acquire (for a price) a chatbot that would read years’ worth of our writing, all the books we’d struggled through, and even our oral presentations and casual conversations. It would know us as well as a parent or sibling. Maybe better.

It would also have a physical appearance. In one of my 1980s novels, the characters often use their computer screens to consult AI programs (I called them “turings”) with names and distinctive appearances and personalities. They get friendly, wise advice that they don’t always want to hear. Similarly, in the world of Philip Pullman’s His Dark Materials, the characters are born with “daemons” in the form of animals. Each daemon is part of its character’s soul; people and their daemons may quarrel with one another, but they suffer when separated.

So after a few years of school, we might graduate to a chat-daemon we could see on a screen and hear on an earbud. It would know us well — and know the world we’re going to live in. It would encourage us, console us, criticize us and nag us when we, not it, had to use some skill instead of passing it off to the chat-daemon.

The menace of the hacked daemon

So educated and so equipped, would our children be happier and more successful than we have been? Much would still depend on the quality of their education and on that of their chat-daemons. A miseducated daemon could ruin its person, and a bad person could exploit its daemon while doing real harm. (And chat-daemons would have to be carefully defended against hackers and other malefactors).

If nothing else, filling kids’ minds with good prose and poetry and mathematics (whether by a teacher or a chat-daemon) would accustom them to thinking with richer vocabularies and more careful logic than most of the people educated under the present system. Our chat-daemons would be in touch with one another, no doubt, and able to call for help when we got into real trouble.

Writing and other technological “advances” have made us what we are today — powerful, but also powerfully dangerous to ourselves and our world. If we can just think through the implications of ChatGPT, we may create companions and mentors that are not so much demonic as the angels of our better nature.