AI will kill these jobs (but create new ones, too)

While experts agree AI will automate many manual, low-level tasks, it will also free up workers to take on more important jobs, such as project management and AI training and auditing.

shutterstock 2270550899
Shutterstock

Generative artificial intelligence (AI) tools such as ChatGPT are rapidly gaining traction, allowing the technology to be used throughout the enterprise to automate a variety of manual tasks performed by workers today, leading to what's expected to be a major shift in the global job market.

In the US and Europe, as many as 300 million jobs could be threatened by some form of AI, according to a March research note by investment bank Goldman Sachs. Fully two-thirds of US jobs could be partially automated through AI, and up to one in four current work tasks could be completely automated by AI in the US and Europe, according to Goldman Sachs.

In particular, roles that require repetitive data entry, legal administration, careers that involve mathematical skills – even healthcare jobs – will all be impacted by AI’s adoption.

goldman sachs graphic 1 Goldman Sachs

As much as 29% of computer-related job tasks could be automated by AI, as well as 28% of work by healthcare practitioners and technical tasks in that field, Goldman Sachs projected.

Career fields with the highest exposure to AI automation are administrative positions (46%) and tasks in legal (44%) professions. Not surprisingly, jobs less likely to be affected tend to be in physically intensive areas such as construction (6%) and maintenance (4%).

While automation through AI is expected to disrupt an array of career fields, including IT, not all jobs in those areas will be affected equally. For example, in the legal sector, paralegal jobs are likely endangered than attorneys — that's one reason the legal sector’s score is so high, according to Goldman Sachs.

Officials at ManpowerGroup are seeing many of the same trends. “A majority of what we’re seeing now are [effects on] jobs that are more information based,” said Rebecca Croucher, head of North America marketing at the global staffing firm. “So, math, law, doctors —  more of a day-to-day diagnosis [of patients].”

Data processing tasks such as data entry are also “going away,” especially in banking, according to Croucher.

goldman sachs graphic 2 Goldman Sachs

“No longer does someone have to sit there and enter invoices or accounts receivables. Any data coming in like that is now being automated,” she said.

With AI taking on so many tasks in so many fields, organizations will be pressed to upskill or reskill their existing workforce. In fact, half of IT employers are upskilling workers to address staffing challenges, according to a June 8 report from ManpowerGroup. The report claims that emerging technologies such as AI and virtual reality could even help hire and train new employees.

Croucher said the need to retrain employees as AI takes on tasks will hit companies “hard and fast.” “I think there is the cirriculum available, but I don’t think the rubber has hit the road yet. In the US, you’re starting to hear more about apprenticeship programs in areas like cybersecurity,” Croucher said. “The need will happen. And when it hits, it will be a big wave.

Knowledge worker jobs are also threatened by AI.

Jack Berkowitz, chief data officer for HR software and service provider ADP, said companies are now taking longer to hire knowledge workers. For the most part, Berkowitz said, ADP’s data shows major dislocations for workers in three areas: digital marketing, digital advertising, and digital sales.

“...That’s because the previous generation of technology in e-commerce and marketing required people to click buttons to manage the search engine optimization," he said. "Now, the tools are automating that. So, it’s the modern job that's getting disrupted.”

Automating yourself out of a job

Berkowitz noted that in his current role, where he is helping to develop AI and machine learning tools to eliminate manual data analytics tasks, he may well be automating himself out of a job. But he calls the technology "amazing" and is sanguine about his future after 30 years working in product development for ADP.

For more than two years, the company has been using various large language models (LLMs), including GPT 4 and Google Bard, to improve the efficiency of ADPs’ skills graph. That application, among other things, maps in-demand skills to what people should be paid for them and the capabilities of the workforce.

“All the job numbers you see [from ADP] about who’s doing what and in what professions, that’s all actually run by machine learning and AI that we developed two or three years ago. So, all of the job numbers — everything that we publish — has machine learning behind it today,” Berkowitz said. “The technology is amazing.

“These large language models turn out to be really good at summarizing data. So, do I need someone to spend 100 hours building a dashboard for me, or will the new tools allow me to do it almost dynamically?” he said. “There are research assistants, associates; there are people who configure business rules around companies. Well, here’s our policy and now we’re going to configure and put those policies into action by either configuring software or training people on processes.

"These systems are really good at reading a policy document and taking action as a result of that,” Berkowitz said.

The good news: worker displacement from automation in the past has usually been offset by new job creation. That trend accounts for the vast majority of long-term employment growth.

Even as ADP automated digital tasks once done by workers, it didn’t reduce headcount as a result. In fact, Berkowitz said, the company ended up paying those data analysts more because those employees could develop new skills and advance up the career ladder.

“So, we literally changed the jobs for 10 or 12 people internally. I think eventually, my job might go away. I think if we do a good enough job, it will. But that’s OK. What’s wrong if I put myself out of work? That means I’ve done something meaningful and contributed,” Berkowitz said. “I’ve done my job. I’m sure they’d find something else for me to do.”

And, he noted, AI is not perfect. For at least the foreseeable future, it will continue to be more of a co-pilot than pilot in charge; people will still be needed to ensure the work produced by generative AI tools is accurate.

A coming productivity boom?

The combination of significant labor cost savings, new job creation, and higher productivity for non-displaced workers raises the possibility of a productivity boom that could boost economic growth substantially, “although the timing of such a boom is hard to predict,” Goldman Sach said.

A study released this week by Upwork, a freelance staffing platform provider, underlines that possibility; it indicates companies plan to hire more people even with the rise of generative AI.

Contrary to the assumption that AI will replace millions of jobs, Upwork’s survey of 1,400 US business leaders across various industries found 64% of C-suite executives plan to hire more professionals of all types due to generative AI. And 49% of all respondents, which included senior managers through C-suite level, plan to hire more freelancers and full-time employees.

Remote-first enterprises are more likely to embrace generative AI, with 68% of full-time remote companies saying they're actively doing so, compared to 53% of companies that are full-time in the office.

Almost six in 10 (59%) business leaders also said they're personally embracing generative AI tools.

The survey, however, found a disconnect between leaders and their teams. Seventy-three percent of C-suite executives think their company is actively using AI tools such as ChatGPT and Midjourney, but just 53% of vice presidents, directors and senior managers say that's the case. Kelly Monahan, managing director of Upwork Research Institute, said some teams are reticent to rollout AI due to a lack of understanding and training around the technology.

“Businesses that want to bridge this gap should roll out a change management strategy that includes communicating the expected outcomes for their workforce in relation to generative AI adoption, and clear policies,” Monahan said in a statement. “It’s important that leaders address fear and uncertainty, and, perhaps most critically, encourage their teams to adopt a learning orientation.”

The arrival of the 'AI age'

Just as the industrial revolution ushered in the industrial age and the computer revolution launched the information age, the AI revolution heralds the coming AI age, according to Anant Agarwal, founder of online education provider edX. Not surprisingly, Agarwal sees the AI age as an opportunity to upskill workers for jobs that will be created through automation.

“With a billion jobs worldwide projected to undergo a dramatic evolution within the next decade due to AI, the demand for rapidly-growing skills, such as AI/machine learning, cloud computing, cybersecurity, product management, project management, and digital social media, is soaring, Agarwal said. Those skill sets, which opened up over five million job opportunities in 2022 alone, “represent a seismic shift in the job market.

“AI will become an augmentation skill that everyone will nee,d no matter what their job. For example, call center specialist to ChatGPT-augmented call center specialist,” Agarwal said.

AI itself will need to be tailored for specific tasks, managed and monitored. For example, prompt engineers will be needed to craft and optimize text prompts for LLMs to achieve desired outcomes. “[It] helps LLMs for rapid iteration in product prototyping and exploration, as it tailors the LLM to better align with the task definition quickly and easily,” said Marshall Choy, senior vice president of product at SambaNova Systems, a Silicon Valley startup that makes semiconductors for AI.

AI systems will also have to be monitored to ensure their output is accurate and useful; that will be the job of an AI auditor.  

Overall, AI competence will be prized by employers, with AI skills, from coding to working alongside AI, understanding its implications, and integrating it into existing structures, necessary in coming months.

“The future belongs to those who are ready to embrace the AI wave,” Agarwal said. “Remember, AI may not take away your job, but could be something that can help you do it more effectively.”

Copyright © 2023 IDG Communications, Inc.

It’s time to break the ChatGPT habit