The big technology news everyone is talking about these days is the booming popularity and potential of ChatGPT, the artificial intelligence "chatbot" prototype released in November by OpenAI, the research organization founded in 2015 by Elon Musk (who is no longer involved) and other top Silicon Valley entrepreneurs.
Ironically, the impetus for OpenAI came partly out of concerns regarding the long-term dangers to humanity presented by artificial intelligence.
ChatGPT is an example of generative AI — algorithms that create new content in response to a user suggesting a specific topic or theme. It has repeatedly shown the ability to generate realistic, humanlike text and write and debug computer programs, among other tasks classically considered creative knowledge work.
The implications regarding the automation of knowledge work are startling. Many observers have focused on the impact on education, as students tasked with writing essays can use it to cheat. But it will have an equally disruptive effect on the business world.
To be clear, researchers are nowhere close to true autonomous artificial intelligence. And as Yann LeCun, Meta's chief AI scientist, said the other day: "In terms of underlying techniques, ChatGPT is not particularly innovative."
Furthermore, ChatGPT cannot be relied on to consistently create accurate or useful results. It can take several versions for its output to be useful; the results may be clichéd; and it sporadically invents information, including citations to nonexistent sources.
ChatGPT may not be able to, now or in the future, automate true creativity. But the majority of writing — and coding — is relatively tactical, and ChatGPT does this starltingly well most of the time.
ChatGPT has the potential to substantially automate much maintenance coding, and turn business content creation in both B2B and B2C settings into an editing process, rather than the writing of new content. My guesstimate is that an editor of ChatGPT content will be five to 10 times as efficient as a writer creating new content. The human role will still be critical, but most of the jobs will be eliminated through automation.
On Monday, Microsoft announced it will invest $10 billion into OpenAI, on top of $3 billion invested since 2019. It also has announced its intent to embed ChatGPT across its suite of products.
Prepare for ChatGPT to be widely adapted quickly, as soon as nine to 18 months. It will have major disruptive effects on knowledge workers globally. Aside from text recognition, which took decades to mature, this will be the biggest AI technology yet.
ChatGPT is free to users, but quite expensive, as it is highly computationally intensive, costing four to eight times more than a Google search. How will companies offering it pay for it? Most likely through monetizing (and eroding) what's left of our privacy.
Advanced economies have been experiencing a tidal wave of disruption since the rise of e-commerce, which sped up with the advent of social networking. The automation of massive amounts of knowledge work through AI comes at a time when societies — and workers — are already exhausted by the pace of change.
Again, this will have an extremely disruptive effect on the knowledge worker class globally, plausibly resulting in the loss of millions of jobs in the next several years. It will play out socially and ideologically in ways we don't yet understand.
Isaac Cheifetz, a Twin Cities executive recruiter, can be reached through catalytic1.com.
