AI can improve some skills but worsen some existing ones

Recent studies have shown that integrating ChatGPT (and likely other chatbots) into knowledge work can increase productivity.

As reported here, Professor Ethan Mollick of the Wharton School took part in a study with many other social scientists and consultants at the Boston Consulting Group (BCG) to see if the tool increased their performance.

“For 18 different tasks selected to be realistic samples of the kinds of work done at [BCG], consultants using ChatGPT-4 outperformed those who did not by a lot”, Mollick wrote in a blog post. “Consultants using AI finished 12.2% more tasks on average, completed tasks 25.1% more quickly, and produced 40% higher quality results than those without”.

The researchers randomly assigned the use of AI to hundreds of consultants to determine the true impact of AI on knowledge work. The same writing, marketing, analytical, persuasive, and creative activities were given to the AI-enabled group and a control group that did not have AI. They included anything from completing market segmentation to producing a press release.

AI and creativity

According to the research, consultants who had access to AI performed noticeably better. This was true for all measurements, regardless of how long it took to complete activities, how many tasks were completed in total, or how well the results were produced.

The role of AI in skill leveling is another important discovery. The highest improvement in performance (43%) was shown in the consultants who began the exercise with the lowest score. Still, but to a lesser extent, the top consultants received a lift. This shows that AI can assist in raising less talented individuals up to the level of top performers.

These results support the conclusions of a related study conducted by Stanford and MIT in the spring. In that study, the performance of 5,000 customer care representatives at a Fortune 500 enterprise software company over the course of a year was examined. The average productivity of agents using AI was found to be 14% higher than that of those without it, with the least competent employees benefiting the most as they were able to finish their work 35% faster.

The entry of AI into the workplace, however, had little to no positive impact on the most skilled workers.

Together, these research findings may have additional effects on the workforce. For instance, companies could discover that they can generate more income with the same employees by achieving more. A workforce with a wider range of skills would develop if highly trained individuals concentrated on more specialized tasks that AI could not complete.

On the other side, increased productivity and efficiency brought on by AI augmentation may potentially raise performance standards, leading to stress or work dissatisfaction for individuals. And this development might result in layoffs in some departments.

AI and jobs

This fact is no longer hypothetical; as evidenced by a recent analysis from job placement company Indeed that looked at job listings and skills, AI is clearly having a genuine impact on the workforce. An in-depth analysis of how future AI will affect employment and the abilities required to accomplish them is provided in an “AI at Work” report.

To determine the exposure level, the report examined 2,600 job skills and more than 55 million job postings on Indeed. About 20% of jobs were deemed to be “extremely” vulnerable to the effects of future AI. Extremely exposed denotes that at least 80% of the job-specific abilities could be performed by AI. A further 45% of job postings are moderately exposed, indicating that between 50% and fewer than 80% of the necessary abilities can be performed by AI.

AI and complex skills

This study demonstrates how AI may assist humans in producing better work and carrying out a variety of jobs. But can AI handle more difficult jobs? To find out, Section School, a startup that specializes in educating people about how to use AI effectively, recently conducted a thought experiment on the analytical capabilities of existing chatbots.

“Before our most recent board meeting, we asked four AI chatbots to give us feedback on our board slide deck”, Section School reported. The quality of output varied across the chatbots, “but Claude [from Anthropic] was almost as good as our human board. It understood the macroeconomic environment, was appropriately ambitious, and quickly got to third-level implications and big-picture opportunities”.

This suggests that in the future, AI advisers may supplement or possibly completely take over the function of human specialists and advisors in assessing complicated judgments, tactics, and plans. Whether AI will actually take the place of human creativity and strategic thinking is still up for debate.

Pros and cons of AI efficiency

While the findings of these multiple studies indicate that AI is beneficial for the workplace, a second paper that examined the effectiveness of job recruiters discovered that those who employed high-quality AI developed laziness, carelessness, and a decreased capacity for independent judgment. Humans have no need to put in extra effort or focus when AI is so good.

The paper reported: “As AI quality increases, humans have fewer incentives to exert effort and remain attentive, allowing the AI to substitute rather than augment their performance”. People let AI take over instead of using it as a tool, essentially falling asleep. This demonstrates that people could readily become overdependent on an AI and fail to exercise their own judgment.

In the not-too-distant future, people may complete their work on autopilot, just like our cars do today. Similar findings from earlier smartphone usage research were reported by The Wall Street Journal. According to that study, as the brain becomes more reliant on phone technology, intelligence declines. The same probably applies to any sort of information technology where we receive stuff without exerting any effort to research, understand, or find it for ourselves.

If so, AI, which increasingly delivers content customized to our unique needs and interests, especially at work, could lead to a reliance that impairs our intelligence. “I worry that human abilities may atrophy”, said Daniel Weld, a professor of human-computer interaction at the University of Washington, in an Axios article.

Human motivation and creativity include comparing ourselves to others and other members of our environment. In many fields, AI has recently shown itself to be more skilled than humans. Therefore, people may be inclined to give up trying if they cannot compete.

A balance between humans and AI at work

The impact of AI on the workforce is becoming clear. According to recent studies, AI-assisted consultants outperform their competitors, greatly increasing output and quality. While AI can help lesser performers catch up with competence, it can also encourage dependency and human deskilling.

The efficiency of AI is a double-edged sword that becomes increasingly obvious as it becomes more integrated into professional functions. Companies must proceed cautiously, using AI’s benefits without sacrificing human abilities and judgment.

Overall, it is evident that the careful application of AI can increase productivity, but companies must take care not to exploit the technology in ways that limit human talent. Maximizing human engagement and leveraging each person’s talents require finding the best way to divide work between humans and AI.

In conclusion, AI tools like ChatGPT hold great promise for improving how we find information, gain explanations, and learn new skills. Someone who previously struggled to accomplish certain tasks can use AI to quickly gain competence. However, over-reliance on AI risks reducing human abilities over time. Workers who let AI do their entire job may appear highly productive but lose motivation and skills. Companies should encourage using AI for learning and augmentation, but not complete substitution. With the right balance, AI can help less experienced people rapidly gain capabilities while more skilled workers further advance their talents. But if AI is exploited to do entire jobs, it could create discrepancies where less competent staff get chosen over more skilled human workers. The key is using AI’s strengths for knowledge enhancement while retaining engaged human judgment, effort, and discretion to avoid deskilling. With responsible use, AI presents exciting potential to both boost productivity and enrich human skills.