How Much More Productive Could Workers Become With AI?
Artificial Intelligence & Machine Learning
Worker productivity soared in the latter half of the twentieth century. By 2000, the average employee in the U.S. was getting around 2.5 times as much done per week as in 1950, leading incomes to rise significantly and the middle class to emerge as a powerful political force.
The main driving force for those improvements was intensive capital investment and the computer revolution. New tools meant that the average output per hour worked soared.
Most people took the innovations of the second half of the twentieth century for granted. Commentators quickly got used to the idea of unlimited economic growth, believing the economy would continue to expand at a prodigious pace of 3 to 4 percent per year. However, after 2000, the pace slowed continually. The gains seen in previous decades reached maturity and the wealth explosion started to run out of steam. Technology continued to develop in the background, but improvements appeared less economically valuable. Wages stagnated, and populism thrived.
Now, though, technology is once again coming to the rescue. After two decades of relative stagnation, a new crop of innovations looks likely to continue driving economic progress and raising productivity for all.
Chief among these is AI, perhaps the most powerful technology humans have ever invented. Intelligence on a machine is enabling workers to outsource cognitive loads to computers, effectively increasing their ability to get tasks done by orders of magnitude.
Working With AI: The Potential Benefits
Estimates suggest that workers that use AI in customer support roles are around 13.8% more productive than their non-AI-assisted counterparts. Data suggest that artificial intelligence can help them solve more queries per hour, helping to reduce call queue times and business expenses. Machines enabled workers to find relevant answers faster and explain them in ways customers could understand and use.
But this example is just the tip of the iceberg. More general intelligence is now opening up the possibility of machines completing vast arrays of highly cognitive tasks that would previously require strenuous human effort.
Plus, an AI productivity company, demonstrates this concept in practice with its AI presentation maker. Like ChatGPT, the tool requires little more than a short written prompt to create a full slide deck for any presentation imaginable. The firm calls it the best generative AI tool for Google Slides because of its ability to mimic hours of human work in seconds.
Naturally, it’s straightforward to use. Plus says that all users need to do is install the Google Slides add-on and then start generating custom presentations in minutes. “Just tell us what kind of presentation you want to make,” the firm boldly tells its website visitors, “and we’ll make it yours.”
Plus believes that the new technology has the potential to supercharge worker productivity. Constructing presentations, documents, and even designs now takes a matter of minutes with new tools, potentially freeing up dozens of hours per week. Workers could conceivably get twice as much work done between Monday and Friday relying on these systems.
Of course, Plus isn’t the only company working on high-productivity applications of AI. Practically every SaaS solution intends to include a combination of natural language processing, deep learning, and large language models in its software. Agent-based systems of the future might be able to schedule meetings, automatically chase up customers, and auto-respond to FAQs and emails, without requiring any labor.
The Limits To Productivity
Of course, AI is not the same as other innovations, like the telephone or fax machine, that increased productivity. Those machines required human input to perform roles. AI, though, is the first example of a technology in history that is able to direct its own activities.
Today, agent-based models are limited, but some examples already exist. It will only be a matter of time before software engineers combine the reasoning abilities of large language models with an objective-driven program to enable AIs to make decisions in the workplace.
Whether this development will benefit wages is unclear. Robots assisting humans in existing roles could increase productivity tremendously. However, those returns might not go to the workers cooperating with them. That’s because AI isn’t merely enhancing work – it’s replacing it wholesale.
The first sectors to feel the AI burn will be those in the knowledge industry. Workers will increasingly find AI taking over more of their roles (and doing a better job of it). Eventually, companies will wonder why they have various people on their payroll and whether they should let them go. The drive for cost-efficiency could make a significant chunk of the white-collar workforce unemployed, forcing them to find new opportunities elsewhere.
AI could also infiltrate into the C-suite, potentially usurping decision-makers and replacing them with algorithms. Once an AI company demonstrates that artificial intelligence generates better CEOs, many other companies will follow, and those right at the top of the income pyramid could find themselves out of a job.
What Will Become More Productive?
Historically, economists thought of capital and labor as being the factors of production that became more productive over time. Better technology embodied in each improved their output, enabling the economic growth of the last two centuries.
But in an AI-powered world, it’s not clear whether speaking of the traditional factors of production makes sense. If AI does all economically-relevant tasks at a lower cost than people, then employee productivity might cease to be a meaningful concept. What will matter will be the capacity and intelligence of the system to carry out plans that enable human well-being. In the short run, AI has vast potential to assist lawyers, legal officials, police officers, and marketers. But, long term, it could price them out of the market entirely.
As such, the direction AI will take the job market remains unclear. Anything short of general intelligence will likely assist workers in specific tasks and enable them to get more done in record time. But anything more than that will ultimately replace them and lead to a fundamentally different kind of world. Suddenly, economics got very interesting indeed.
