Asking Nikkei Readers: Generative AI is "Useful for Work" 75% — A Competent Assistant, Reduces Human Errors

• BtoB
• Report

Interest in generative AI is growing every day since the launch and widespread use of ChatGPT, an interactive AI chatbot developed by the company OpenAI. In a survey conducted toward Nikkei readers through April 3rd-6th, more than 70% said they are interested in the technology, while 20% said they already use it at work.

What kind of applications do people in business use AI for? What kind of expectations and issues do they have? Let's take a look deeper into the results of the survey.

The survey was conducted online, targeting the registered Nikkei ID members (subscribers of Nikkei's media). A total of 2,099 respondents who are employed participated.

1. "Use it for research" 57% — good enough for supplemental work

What do you use generative AI for?

When we asked the already users of generative AI about what they use them for, the most common answer was "research" (57%), followed by "gathering ideas" (51%) and "creating texts" (48%). "Programming/coding assistance" was 16%.

"Research" was used across broad industries, from manufacturing such as Automobile/Transportation equipment (56%) and Machinery/Heavy electrical equipment (56%), to Wholesale/Retail/Commerce (64%), Finance/Securities/Insurance (59%), and Healthcare (64%). Utilization in "programming assistance" was spreading in Information processing/SI/Software (29%) and Machinery/Heavy electrical equipment (22%).

Regarding the convenience and issues when using, most respondents felt that the accuracy of the information was insufficient and they were negative about relying entirely on AI, but on the other hand, many respondents thought it is good enough for supplementary tasks.

Another challenging issue pointed out was "how to ask questions to get better answers" (Automobile/Transportation equipment, Contract employee). In fact, while using ChatGPT to gather information, I actually experienced that the answers depend on how we put the question. Perhaps many people think it is only a matter of time before the AI's accuracy improves, and therefore, advancing their own skills for better utilization is the matter on the human side.

2. Effective use case: "Translation into other languages" 58%

In what situations are generative AI effective to use?

In the survey, respondents were asked, "In what situations are generative AI effective to use?" The most common response was "making research more efficient" (76%). That was especially high in industries of Telecommunication services (84%), Consulting/Accounting/Legal (80%), and Machinery/Heavy electrical equipment (80%).

"Translation into other languages" (58%) was the next most common use case. By industry, Telecommunication services (66%) and Wholesale/Retail/Commerce (63%) were high. From an open-ended question, there were a notable amount of responses praising the precision of creating English texts.

3. AI as a "companion" to humans. Some call the importance of the training to use

Will the widespread of generative AI help your work?

When asked whether the widespread of generative AI helps their work, 25% of the respondents answered: "strongly agree." When combined with the "agree" (50%), three-quarters of the total were positive. The reasons are as follows.

Generative AI can improve a company's competitiveness, even if it only reduces human errors. When the accuracy of information is improved, it will spread even more, instantly.

On the other hand, many people at work feel they need to redefine the "role of humans," facing the wake of the AI era. There were some opinions like the following in the survey.

The conspiracy of "AI threat" taking away people's jobs is rumoring about, but it is impossible to choose a direction of our path from just AI-generated information, and without human creativity. As the birthrate declines and the population ages, the coexistence of AI and humans will become a top social issue. The survey results reaffirm the necessity to have further debate on this issue.