Artificial intelligence promises to revolutionize the workplace, but not many professionals know how to best utilize these tools effectively. New research led by François Candelon, managing director at the Boston Consulting Group (BCG), provides some valuable answers to this question.
Candelon collaborated with leading scholars from Harvard, MIT, and Wharton to conduct a large-scale experiment examining how over 750 BCG consultants performed real-world tasks with and without the assistance of generative AI. The study uncovered valuable insights on when and how AI tools can boost productivity and creativity at work.
During an interview with the Wall Street Journal, Candelon shared what he learned from this research. Most notably, he explained that generative AI works best for creative tasks but can actually hinder analytical problem-solving. Additionally, lower performers benefited the most from AI assistance, indicating major implications for companies’ people strategies going forward.
The Unique Value of This Scientific Approach
What set this study apart was its academic rigor and real-world applicability. Consultants completed tasks drawn from actual BCG client projects across two categories: creative product innovation and problem-solving. Control groups enabled randomized testing to precisely isolate the impact of AI assistance.
“We used very advanced techniques as well, for instance, on semantics, to be able to see the similarity of ideas, the difference of them, and so on,” Candelon told the WSJ. “We worked extensively and intensively with scholars to make sure that we were doing it the right way.”
Key Insight #1 – AI Boosts Creativity More Than Critical Thinking
On creative tasks, consultants generated more novel, higher-quality ideas with the help of artificial intelligence. However, for analytical problem-solving, the technological tool actually reduced performance.
“AI was able to persuade humans, who are well known for their critical thinking… of something that was wrong,” Candelon explained. This demonstrates the double-edged nature of generative AI – while it can enhance creativity, it may provide a dangerous sense of comfort and the blind acceptance of ideas that may come from an inaccurate and untrustworthy source.
Key Insight #2 – Low Performers Benefit the Most from AI
The experiment also showed that less skilled consultants improved far more with AI assistance compared to top performers.
“The ones [consultants] who are below average [in performance] were actually benefiting much more from AI, which means that for a company, it has plenty of implications for their people strategy,” said Candelon.
In other words, AI could provide a leveling effect, boosting productivity across the board. However, it may be especially impactful for less experienced employees who may become more efficient at certain tasks by using the tool.
Key Insight #3 – AI Changes How Humans and Machines Can Best Collaborate
Rather than simply automating tasks, artificial intelligence requires rethinking workflows from the ground up. Effective human-AI collaboration is critical.
As one example, Candelon suggested that instead of starting ideation with AI, “maybe you should do it the other way around. Start with humans, with the diversity [of ideas], and then try to see how to improve these ideas with AI.”
Also read: Best Free AI Content Generator: Top 10 for February 2024
What Companies and Professionals Need to Know
What do these findings mean for harnessing AI’s potential? Candelon emphasized several implications:
First, data infrastructure becomes even more vital. Organizations must curate quality datasets to train their AI tools in a way that they can be put to good use for their specific needs.
Additionally, generative AI requires a complete re-engineering of functions and processes with human-machine collaboration in mind. Rather than focusing on specific use cases, professionals should consider how AI can help solve high-level business problems related to efficiency, cost reduction, innovation, and more.
Finally, because AI’s performance differs drastically depending on context, real-world experimentation is critical. Companies must continually test the tool to understand where artificial intelligence creates value versus areas where it may undermine critical thinking.
“It’s difficult for people to understand [whether AI] is providing value or not”, Candelon noted. “For companies, it is important to understand where you can use it, where you cannot use it, by experimenting”.
These rigorous findings provide additional insights into the best practices for leveraging AI, boosting productivity, enhancing creativity, and rethinking strategies when it comes to hiring talent. While artificial intelligence holds great promise, realizing its full potential requires a thoughtful implementation process tailored to specific tasks and contexts.
Companies must continually experiment and evaluate if these tools are truly enhancing performance versus just persuading employees with seemingly credible but inaccurate information. By following these guidelines, both companies and professionals can thrive amidst the AI revolution.
Maximizing AI’s Potential at Work: 6 More Key Tips from the Experts
While AI holds incredible promise for increasing workplace productivity, realizing the full benefits requires thoughtful strategy. Experts share these 6 key steps that professionals may take to integrate artificial intelligence successfully to enhance their output:
#1 – Improving Efficiency with Task Automation
AI writing tools can automate drafting everything from reports to slide decks, providing templates for humans to refine. This facilitates faster content creation versus starting from a blank page. However, critical review is still essential, as AI can (and often does) generate factually inaccurate content pieces.
Also read: OpenAI’s Text-to-Video AI Model Sora Is About to Wipe Out Entire Industries: Here’s Why
#2 – Focusing Questions to Optimize Relevance
Rather than taking a Google-like search approach, professionals should clearly frame queries around achieving their desired outcomes. The more precisely the problem and goals are defined, the higher the odds that the AI model can deliver relevant assistance and output.
#3 – Verifying Anything Used Downstream
Before presenting or acting upon any AI-generated information, rigorous verification is non-negotiable, even if just checking for factual inconsistencies or unrealistic suggestions. Blindly following the guidance of an AI tool risks outcomes worse than not being assisted at all.
#4 – Starting with Humans Before Collaborating with AI
For analytical tasks, experts suggest that humans should provide a wide range of ideas and hypotheses that the AI tool can help refine and enhance. This takes advantage of human creativity while using AI for incremental improvements versus generative tasks where AI should contribute first from its relatively limited knowledge base and creative toolkit.
#5 – Asking “Dumb” Questions Without Judgment
Even when humans know the answers to some questions, the anxiety around being perceived as uninformed often prevents them from asking clarifying questions. AI-powered chatbots provide judgment-free opportunities to strengthen one’s understanding of key nuances about any topic. This is especially helpful for those who are less experienced in the subject. But remember to fact-check the AI’s explanations as it can confidently provide wrong information, called “AI hallucinations.”
#6 – Continuous Hands-On Experimentation
Developing an intuitive sense of when and how to use artificial intelligence effectively requires actively testing applications customized to one’s role. While best practices are emerging, optimal utilization strategies remain context-dependent. Professionals should continually try new prompts and workflows to determine what brings the greatest benefit and produces the optimal output for their specific needs.