How generative AI is impacting BCG consultant productivity
BCG decided to find out how generative AI improves the productivity of their employees. And spent study, which was attended by more than 750 BCG consultants around the world. The idea was to figure out which tasks AI helps with and which ones it doesn’t.
What conclusions do we ultimately get:
In creativity tasks where you had to come up with something and conceptualize something (in this case it was necessary to develop ideas for new products and a strategy for entering the market), 90% of consultants showed an increase in productivity. They were 40% more productive than those who did not use GPT in these tasks.
But when subjects used GPT for business analysis tasks, they performed 23% worse than those who performed the task without GPT.
The study demonstrates a paradox: people seem to mistrust technology in areas where it can be of great benefit (creative tasks), and overly trust it in areas where technology is ineffective (analysis, working with facts, expert judgment).
As far as I understand, the higher the expertise and performance of the specialist, the less the effects of using GPT.
Also note that employees who received AI training performed less well than those who did not receive the training. A possible reason is that those who completed the training were more confident in their actions and did not evaluate their results critically enough.
For employees who are not known for creativity and performance, a correlation was found that the more they edited GPT answers, the worse the final result was. So much for AI competing with human creativity (and these are BCG consultants!).
And the final chord, AI increases creativity in individual work, but in a group process significantly reduces the diversity of ideas (this is also taking into account the fact that, according to a number of studies, group brainstorming in general reduces the likelihood of truly original ideas).
