In the 1960s and ’70s, the U.S. manufacturing sector began hemorrhaging jobs as companies outsourced production to places like India and China, where wages were lower and labor was less expensive.
Today, American companies are at the beginning of another wave of outsourcing. This time, the jobs aren’t heading overseas. Instead, they’re being turned over to artificial intelligence.
“For example, if you’ve ever called a customer service line in the past, you’d have another human, most of the time in another country, pick up your call,” said Gaurav Saluja. “But now, these companies are moving them online and having AI pick up the phone the first time. If AI can solve your problem, they don’t have to bring it to a human … saving businesses quite a bit of money.”
Saluja and his research partner, Gaurav Sachdeva, are UWM students studying this new trend of AI outsourcing under the mentorship of Avik Chakrabarti, associate professor of economics. Saluja, a double major in economics and computer science, and Sachdeva, who is majoring in management and finance, have completed their first year at UWM. They presented their research at the UWM Undergraduate Research Symposium in the spring.
The incentive and pitfalls of AI outsourcing
Offloading work onto a digital assistant is becoming an attractive option for businesses looking to cut labor costs. Instead of using employees to take customer service calls or handle data entry, they are increasingly turning to AI. This not only saves money on wages that would have otherwise been paid to the employee, but it decreases overhead costs as well.
“You don’t have to go through the hiring process,” Sachdeva pointed out.
their research poster at the UWM Undergraduate
Research Symposium in April. (Photo courtesy of G
aurav Saluja)
And even if you don’t completely outsource the jobs, AI can still increase a business’s productivity, he added. Say it takes a software engineer 10 hours to write quality computer code for a project. Artificial intelligence can generate that code in two to three hours at a lower standard of quality. The engineer might take two to three hours to edit and improve upon the AI-generated code. A project that originally took 10 hours now takes only six. Now, the engineer is free to work on tasks to grow the business.
But AI has its drawbacks. Many critics worry about its lack of flexibility, Saluja noted.
For example, “social media trends – a human may be able to recognize and capitalize on that and build a lot of revenue for a business, whereas AI may not immediately understand that or be able to predict it,” he said.
There are also concerns around ethics. Companies are beginning to explore replacing higher level managerial positions with AI, Saluja said. While artificial intelligence may be able to optimize profits, it may not be able to address the very human concerns faced by everyday workers.
New job opportunities
One thing that Saluja and Sachdeva noticed during their research was that even as people lost their jobs to AI, new positions were being created.
“Let’s say that 10 software engineers are needed to maintain this database, and let’s say that AI increases their job efficiency by about 200%. Now, I might only need five software engineers to maintain the AI,” he said. “They can either cut those five positions, or they can let those five software engineers now work on building (a new product). Now they’re increasing their productivity, which can increase their competitiveness in the market.”
But that’s little comfort to workers who do lose their jobs to AI software. In many cases, companies find that while they’ve saved on unskilled labor, they now need employees who can work with the AI, perform quality control checks, tweak algorithms and so on. Sachdeva discovered that many of those businesses landed a solution for both problems: They would retrain those “obsolete” workers.
“There are workplaces that are giving employees the opportunity, if their job was replaced, to learn about AI and how it’s used in that particular job. They gain more skills to manage that AI,” said Sachdeva.
Still, many workers will lose their original jobs. Sachdeva acknowledges that reality but thinks that we ought to approach the advent of AI in the same way we did when new technologies like computers or the internet were introduced.
“It’s not going to replace workers; it’s going to change the responsibilities of the workers and what they’re asked to do,” Sachdeva said. “AI is changing so fast that we don’t know where it’s going to lead.”
The next steps
Their research is not over. Saluja and Sachdeva are planning to continue their work with a specific focus on how AI outsourcing is affecting Milwaukee companies. The two plan to approach several businesses in the city, ranging from small businesses to large corporations, to ask questions about how leaders are thinking about artificial intelligence in the workforce and if they plan to use it – or have already implemented it.
“I’m particularly interested in how some companies are using AI for those higher-level manager positions. If we do have any of those in Milwaukee, hopefully, a few months down the line, we’ll have some preliminary results as to how effective AI is at doing those jobs,” Saluja said.
Both are excited to stay involved with research at UWM over the next year.
“Having an undergraduate research project helps you take what you’re learning in your classes and apply it to hands-on experience early on in your college career,” said Sachdeva. “We’re getting that knowledge and learning soft skills as well: Communication, teamwork and being able to manage a project.”
Those are some of the skills that AI can never replace.
Written by Sarah Vickery
Link to original story: https://uwm.edu/news/undergrad-students-study-ai-outsourcing-and-what-it-means-for-human-jobs/