A first-of-its-kind study on AI's effectiveness has reportedly unveiled that people mistrust the tech for the wrong reasons while trusting it for tasks where it might mess up.

The study also found that AI increases performance in creative tasks but performs poorly in problem-solving, proving that adopting generative AI is not without drawbacks. 

The September 2023 study by the Boston Consulting Group reportedly got back in the news this week after one of the study's authors, François Candelon made a discussion about generative AI in the workplace on the Wall Street Journal's Executive Insights podcast, Business Insider reported.

The study reportedly showed that about 90% of the participants improved their performance while employing generative AI for any task involving ideation and content production.

The participants also converged on a performance level that was 40% higher than those working on the identical assignment without using OpenAI's GPT-4. Interestingly, the study found that the most benefits were seen when people did not try to enhance the technology's output suggestions and accepted GPT-4's suggestions as is.

(Photo : Photo by LIONEL BONAVENTURE/AFP via Getty Images)
This picture taken on January 23, 2023 in Toulouse, southwestern France, shows screens displaying the logos of OpenAI and ChatGPT. - ChatGPT is a conversational artificial intelligence software application developed by OpenAI.

AI's Problem Solving

Generative AI effectivity in idea creation should be expected, as AI is currently at the forefront of its capabilities regarding creative ideas. According to the study, participants who used AI for problem-solving tasks performed 23% worse compared to those who did not use the tool.

The study found that when generative AI is used incorrectly for the wrong tasks, it can cause significant value destruction. The research showed that generative AI, which can be a "double-edged sword" with its "relatively uniform output," can also decrease a group's diversity of thought by 41%.

According to Candelon, their goal was to determine how humans and AI could collaborate to benefit businesses. To do that, he enlisted the help of experts from prestigious universities, including MIT, Wharton, Harvard Business School, and the University of Warwick. Also, he asked his consulting company's staff to carry out the experiment.

Assignments, including "creative product innovation," were among the practical activities assigned to the over 750 study participants. According to Candelon, the participants were instructed to utilize the OpenAI tool GPT-4 to assist them with tasks such as developing focus group questions, implementing a successful social media rollout, and selling the shoe concepts to their supervisor.

Read Also: UPenn to Launch AI Degree, Becomes World's First Ivy League School With New Major

AI Implications on the Workforce

Candelon told The Journal that AI is exceedingly powerful and unavoidable. He also told the outlet that "there is this famous quote saying that humans won't get replaced by AI. They will get replaced by humans using AI." 

According to Candelon, the study shows how generative AI will make data even more important in the workplace, requiring people to reevaluate their workflows and identify opportunities for human and AI collaboration.

Related Article: Harvard, BCG Study Sheds Light on Impacts of AI in the Workplace

(Photo: Tech Times)