The fight over the future of AI

Artificial intelligence is the automation of some parts of human thinking: simple analysis, synthesis, aggregation. That represents a sea change The strange thing about ChatGPT is that it may in fact be both underappreciated and overhyped at the same time. — Navneet Alang

When Google search first emerged, it felt like magic. Now, not so much.

Punch in a query — for a restaurant, for a historical fact, or worse, some actual knowledge — and you’ll often get a results page full of sludge, ads and hits that you may or not be looking for.

New developments in artificial intelligence may, however, end up flipping that all on its head.

You could be forgiven for being both weary and wary of “the next big thing” in tech. That refrain has not only become so common as to be almost meaningless, there have also seen a slew of times that such prognostications simply haven’t come to pass.

Yet, now in 2023, it is almost impossible to escape chatter about artificial intelligence, especially ChatGPT, the intelligent chatbot that seems to have the world in a tizzy. Put out by OpenAI, the research company based in San Francisco, ChatGPT has become even more talked about after it was reported that Microsoft is investing $10 billion (U.S.) into OpenAI, in part so it too can have access to the tech everyone is clamouring for.

That investment represents a serious threat to Google. Not only does it represent a significant change for how we think about tech, it also threatens to upend search, how we work — in addition to a whole lot more.

What’s the big deal? It could be as simple as the fact that AI is in fact the next big thing.

Before you ask: yes, there are myriad problems and flaws. And yes, Silicon Valley has yet again unleashed a technology upon the world apparently without the slightest care about what its effects might be.

But the strange thing about ChatGPT is that it may in fact be both underappreciated and overhyped at the same time: that, on the one hand, it is clear it is not even close to being truly intelligent, while also being troubling in about a hundred ways; but on the other hand, it promises to be revolutionary all the same.

The current round of AI essentially works like a smart chatbot — enter a prompt and watch it spit out a response — but one the capabilities of which can be a bit shocking.

Go to OpenAI’s website and enter “make a painting of Toronto’s skyline in the style of an Impressionist” into image generator DALL-E, and the results are actually quite good. Ask ChatGPT any number of things — to write an essay about the history of Chinese restaurants in Canada, to clean up computer code for a simple app, or to write a poem in the style of Shakespeare — and it will do so with reasonable competency.

What differentiates ChatGPT from search or other forms of technology is that it can do a reasonably good job of synthesizing information and presenting it in a far more comprehensive way than a mere search result. Just this morning, I asked ChatGPT to tell me about the ills of English colonialism, and found multiparagraph response detailing the exploitation, prejudice and extraction of wealth to be remarkably complete.

That kind of capability to understand normal human speech and present information in a coherent way means that Google and its search business must react to AI. It’s been speculated that one reason Microsoft is investing in OpenAI is to challenge Google in search by offering up a natural language alternative.

Google have some of the finest engineers on the planet and have pioneered a lot of AI tech, so the fight will be both interesting and evenly matched.

But for its part, Microsoft intends to go further, integrating AI into its Office products. Imagine being able to ask PowerPoint to produce a presentation based on these notes, using a tropical theme, or simply asking Excel to organize data first by year, but then by revenue. These things are quite plausible but also mean that how we work might change quite radically.

The reason is that, quite simply, AI represents the automation of some parts of human thinking: simple analysis, synthesis, aggregation. That represents a sea change — and is something you might fairly call “the next big thing” in not just tech, but society at large.

It is important to note that ChatGPT is riddled with problems. Because it is “trained” on existing data, it can spit out incorrect information (in an authoritative tone). It has also massively complicated the problem of misinformation, not to mention questions of authorship and copyright — issues that the folks at OpenAI either didn’t care enough to consider or simply considered beneath them.

It’s also worth keeping in mind that iPhones and the web and Facebook haven’t magically altered society for the better; one need only look at what is happening on Toronto’s transit system these days to see that our world is far from some magical utopia.

Even with all those caveats, it’s hard not to see the potential in AI. For established tech players like Google, it represents a clear and immediate threat. For the rest of tech, AI will likely start to infiltrate most things. And for the rest of us, we’re simply going to have to figure out how this remarkable new technology might be deployed to help rather than harm.