David Wright is not impressed by the so-called terrifying existential threat posed by artificial intelligence.
Wright is learning to use AI as a tool for his job, providing technical support at Tunxis Community College, in Farmington, Connecticut.. AI is another technology that he needs to keep up with, as IT people do throughout their whole careers.
“If you don't make changes to adapt and take advantage of new technologies, you'll fall behind and ensure you're not successful,” Wright said.
He added, “New technologies won't replace your job if you make them part of your job. Learning them will also protect you from those who would use them against you.”
Wright’s calm approach to planning for the future of coexisting with AI is similar to other workers and experts we interviewed. And it’s an antidote to dire media predictions.
Everybody panic!
For an example of sky-is-falling panic, look to the UK tabloid The Sun, which took a break from publishing bikini photos, celebrity gossip and reality TV updates to issue a warning in late June:: “AI ‘job-pocalypse’ coming as bots could replace ONE BILLION jobs from people—and they will be better than humans.”
Also in June, McKinsey issued a good news/bad news report about the effect of AI on business. The good news: AI could add $2.6 trillion to $4.4 trillion annually to the global economy, significantly more than the entire GDP of the UK in 2021, which was $3.1 trillion.
The bad news: Half of today’s work activities could be automated between 2030 and 2060, McKinsey said.
IBM plans to replace 7,800 jobs with AI and has paused hiring some positions. CEO Arvind Krishna said he thinks 30% of back-office roles could be replaced within five years.
And U.S.-based employers cited AI as the reason for 3,900 job cuts in May, according to a report by Challenger, Gray & Christmas.
It’s not so simple
Reality is likely to be more complicated than AI simply taking the jobs.
Historically, automation has increased employment for the simple reason that when a service gets cheaper and more accessible, people want more of it.
For example, in the fifty-plus years since the introduction of the automated teller machine (ATM), the number of bank tellers increased rather than decreased, according to James Pethokoukis at the American Enterprise Institute. Machines enabled banks to open many smaller branches and put a couple of tellers in each of those branches. The job of a bank teller evolved to focus on sales and customer support.
Similarly, automated scanners increased hires for cashiers, and automated legal discovery increased employment for paralegals, Pethokoukis said.
The trend goes back to the 19th century, as automated textile manufacturing increased demand for weavers, Pethokoukis said.
The advent of the spreadsheet in the 1970s decreased the number of bookkeepers and accounting clerks but increased the number of accountants. Since 1980, 400,000 bookkeeping and accounting clerk jobs disappeared, but 600,000 accounting jobs were added. Accounting got cheaper and faster, and people wanted more of it.
In 2016, legendary computer scientist Geoffrey Hinton, known as the Godfather of AI, predicted that deep learning would be able to replace radiologists by 2021. In reality, the number of radiologists working in the US increased rather than decreased as AI proved unreliable.
Nonetheless, Hinton is now convinced AI is dangerous. He’s quit his job at Google and, in a statement to the New York Times, said he regrets his life’s work. He co-signed a statement warning about the existential threat of AI; other co-signers include luminaries such as OpenAI CEO Sam Altman, US Congressman Ted Lieu, and Bill Gates. (Many of these signatories don’t seem sufficiently scared of AI to stop working on it and getting rich from it.)
Notwithstanding the apocalyptic warnings, the evidence for existential danger from AI isn’t there—not for our jobs or our existence as a species.
Things will be different.
But AI is already transforming the workplace and lives.
For some, AI is more about perception than reality. Matt Heusser, managing director of Excelon Development, gives the same advice about AI tools as he’s given about any software automation: Learn the tools to get the job, but don’t expect to use the tools much on the job. Instead, do the best work you can do.
Excelon provides software testing and development, contracting and recruitment and writes articles on software development.
However, AI tools may well get considerably better over time, Heusser said. “Right now it is much more hype than reality,” he said. The exception is GitHub Copilot, which he described as “amazing.”
“Steve,” a bioinformatics researcher who asked us not to use his real name, said he started using ChatGPT at the urging of his boss, who is enthusiastic about the potential. Steve is more skeptical.
“Some of my co-workers say it's sped up their coding considerably, but for me so far it's a wash,” he said. “I spend so much time getting ChatGPT to understand what I want the code to do, and/or debugging its final output that I might as well have written the whole thing from scratch.”
Wright, the college technical support staffer, noted a similarity to a 1958 children’s book, “Danny Dunn and the Homework Machine.” Set in a small town in 1950s America, the book tells the story of the titular boy hero, who gets access to a computer and programs it to do his homework. Danny learns that to program a computer to do something, you have to know how to do it yourself, and finds that programming the computer turned out to be as much, or more work, than simply doing the studying. (Journalist David Owen wrote about the book and the lessons for AI today in The New Yorker this year.)
Wright said, “For those willing to do it, incorporating AI in their jobs could be like Danny Dunn programming a computer to do his homework. They end up learning more about their jobs and getting better at them while trying to get the AI to do their work for them.”
Thus, AI becomes a tool for job training.
“Steve,” the bioinformatics researcher, said that AI may improve at some point. “I'll keep trying because if it does change—i.e., if the AI becomes able to write good code faster than I can—I want to be ready for it. But I'm not holding my breath.”
Jobs are safe—for now
Karan Batta, Oracle VP of product management, said he expects AI won’t replace jobs in large numbers for at least ten years, if that. AI will make jobs more efficient and optimize work, but it won’t replace jobs. “I don’t think there’s a near-future certainty that somebody will need to be worried about their career,” he said.
However, AI will have a significant impact in the long term, said Holger Mueller, VP and principal analyst, Constellation Research, whose professional focus includes the future of work. “Everybody needs to assess, ‘Can technology replace my job?’ much more than they did before.”
Careers requiring creativity will be valuable in the workplace, said Christian Renaud, global head of Technology, Media and Telecommunications for S&P Global Market Intelligence. Science, technology, and any field that depends on flashes of insights and making seemingly unrelated connections will be increasingly valuable in the AI era,
“Those kinds of cognitive leaps and interdisciplinary leaps are not something that artificial intelligence is well equipped to do because so much of that requires that human spark,” Renaud said. You don’t need to be a genius, just an intelligent person with extensive experience and interdisciplinary knowledge.
On the other hand, careers that require collating, retaining, and retrieving massive amounts of information—like legal research—are more likely to be automated, Renaud said.
Mueller agreed that creativity will be important and said technology skills will also be essential. “Be at the center of technology, because that can make the rest of your career attractive,” he said. The good news is that working with technology requires less education than in the past; low-code/no-code tools can enable people to work without math or computer science degrees. AI tools such as Microsoft Power Apps make it easier for people to create software simply by speaking.
However, understanding business has become more important than technical skills, Mueller said.
AI will dominate rote work, repetitive, routine jobs like approving invoices, which may even disappear, said Patrick Moorhead, founder, CEO and chief analyst for Moor Insights & Strategy. “If you’re sitting at your desk and pieces of paper come in on the left side, and you check them and put them in a bin on the right side for somebody to process for payment, you want to consider another career,” he said.
User jrprov1 on Reddit put it this way: “What I see from AI for sales, to date, is the ability to do more dumb things faster. That will change in the years ahead, but right now B2B salespeople have nothing to worry about.” The Reddit user said they have been head of sales and marketing at B2B technology companies for decades.
People looking to continue to thrive in their careers in the face of AI should look to provide more value. For example, accountants should become auditors. “Go hardcore,” Moorhead said.
RELATED: Nvidia, ServiceNow team to bring generative AI to enterprises