ChatGPT improves quality of work, research finds
Research from Harvard Business School involving consultants from Boston Consulting Group has found workers are up to 40% more productive and 25% faster at work when using ChatGPT.
ChatGPT is a form of generative artificial intelligence (AI), a tool that allows users to enter prompts for the creation of content including images, text and videos.
The joint study between Harvard Business School and international professional services firm Boston Consulting Group (BCG) has provided an early glimpse into the potential for large language models and the benefits to adopting AI.
The study involved 758 BCG consultants last year and gave them various tasks to complete, some with access to AI and others without.
The quality and speed of the work undertaken by the groups was then compared.
The study revealed that those trained in using LLMs performed better than those left to themselves to learn, but overall those using ChatGPT had a marked improvement in quality and speed of work delivered, with the tool found to improve the quality of work by up to 40%. It made employees about 25% faster.
BCG New Zealand co-managing partner Kelly Newton said the study also showed the software helped underperformers the most, narrowing the gap in ability between consultants.
It was not clear why, but Newton said the reason for this could be that the AI filled gaps that held some back, such as the ability to articulate their ideas clearly, to think strategically and develop action plans, or counteract some procrastination.
This was yet another reason for businesses and their workforces to introduce such technology to their operations, she said.
“I was shocked personally by the results from the study in terms of the level of productivity that it claims,” said Newton. “Companies have recognised that this is something new and we should figure out how to harness it for our benefit.
“It’s imperative that businesses start to experiment and they allow people some flexibility to try AI, but do it in a way where they are mindful of data.”
While proven to be a beneficial tool that could be used create content and plans, she said AI had limitations and the success in its use was about the execution and implementation of those plans — by humans.
She said she believed ChatGPT and other generative AIs would increasingly be introduced as a tool to boost productivity in the workplace over time, similarly to how there had been productivity gains from the introduction of computers and internet use, or data processing visualisation tools such as Alteryx and Tableau, which had become common aids for work.
“Employers need to keep the growth objective in mind, and not be tempted to see AIs as a cost-cutting measure,” said Newton.
“The type of work might change, but it’s not going to replace people. That’s why you see so much emphasis on encouraging workforces to adopt and learn the skills and to experiment with it because it will be a part of the future workforce.”
Workplaces that had successfully adopted AI already typically dedicated 10% of their AI effort to algorithms, 20% to data and technological, and 70% to business and people transformation, BCG said.
“Just as companies will sit down with new employees to teach them their other computer systems, I think it’s likely before too long that most companies have a course teaching new starters to use ChatGPT or other AIs, and how to use them ethically.”