On September 21, the Positive AI Community gathered around a webinar on the impact of generative AI, which brought together more than 230 employees from the association’s member companies. On this occasion, François Candelon, Senior Associate Director and World Director of the BCG Henderson Institute – BCG think tank – presented the results of the recent BCG study on the interest and limits of the use of generative AI in work.

Almost a year after the birth of ChatGPT, the adoption of this new technology in the world of work calls for in-depth reflection for business leaders.

“We see that AI combined with humans is extremely efficient on the creative side”, François Candelon, BCG

First lesson: when AI is used wisely, more than 90% of participants improved their performance with GPT-4 for creative ideation purposes (for example: improving the design of a product). They also converged on a 40% higher level of performance compared to those working on the same task without the tool.

But the results nevertheless describe a paradox: although ChatGPT manages to find more in-depth ideas, and therefore judged to be better, the diversity of ideas is, for its part, less important than those imagined by humans. Around 70% of participants believe that heavy use of Chat GPT-4 can stifle their creative abilities over time.

Finally, if consultants seem to be wary of AI in areas where it can bring real added value, conversely, they have tended to trust it too much in those relating to more complex tasks (for example: defining a business strategy), which is not part of the tool’s current skills. Thus, those who used GPT-4 for a task outside the scope it addresses, obtained worse results (-23%) than those who did not use the tool.

The need for change management within companies

According to the conclusions of the study, these initial results seem to indicate that it is more necessary than ever for companies to implement an internal change support strategy. And these decisions will not be made without the key role of business leaders. Indeed, they must be led to precisely think about the tasks that can benefit or be damaged by generative AI within their organization.

Several action levers were notably identified by the authors of the study:

Implementing a data strategy to remain competitive,
Continued testing and experimentation of AI tools as technology evolves
Implementation of an HR and management strategy to identify the new AI professions involved (such as Prompt Engineers), cultivate “human-AI” complementarity and promote learning of these tools internally.
And perhaps most importantly, leaders must continually review their decisions as the AI skill frontier advances.

To read the BCG study in its entirety

The Positive AI Community, a community of exchanges for members of the association

This event – like those to be organized soon – is in line with the association’s commitments, which aim to create a space for exchange and sharing of good practices. The Positive AI Community aims to become the reference community on all ethical AI subjects to progress collectively on the application of responsible artificial intelligence within organizations.

Would you like to join the Positive initiative and benefit from upcoming Positive AI Community events? Join us !