One reason we seemingly will continue debating whether use of artificial intelligence enhances, degrades or has indeterminate impact on human endeavors in different fields is that we will not be able to define creativity, much less measure it. What is original or “novel?” Does usefulness matter? How would you measure it?
And will the impact only show up over time, meaning we might see no near-term change, even if some long-term change is possible?
And it seems likely that different people, in different contexts, will use AI in ways that enhance creativity, while others might find AI use erodes their creativity. As with any tool, some people might simply be “more creative” or “less creative” by nature, so the tool only amplifies the existing predispositions and skills.
Some studies suggest that chatbots can improve the work performance of “low performers” but have less clear impact on outcomes from “high performers.”
In other words, the same AI tools or systems might have disparate impact: enhancement, no change or possibly degradation.
Studies on AI Systems' Impact on Worker Productivity by Performance Level |
Study Title | Publication Date | Publisher/Journal | Key Conclusions |
"The Complementarity Between Artificial Intelligence and Human Labor: A Difference-in-Differences Analysis of Healthcare Workers" | 2023 | Management Science | AI-assisted diagnosis tools provided greater productivity gains for less experienced physicians (30-35% improvement) compared to highly experienced ones (10-15% improvement). |
"Artificial Intelligence and the Modern Productivity Paradox" | 2022 | NBER Working Paper | Found that AI adoption initially widened productivity gaps but eventually helped low performers catch up; suggested organizational learning curves as a critical factor. |
"Unequal Effects: Examining the Differential Impact of AI Tools on Worker Productivity" | 2023 | Organization Science | Workers in the bottom performance quartile showed 45% productivity improvement with AI assistance vs. 12% for top quartile performers. Authors suggest skill substitution vs. complementarity as explanation. |
"AI at Work: The Effects of AI Tools on Knowledge Worker Performance" | 2022 | MIS Quarterly | Lower-skilled workers showed 32% improvement in output quality with AI writing tools, while high performers showed only 8% improvement but reported using AI primarily for routine tasks. |
"Leveling the Playing Field? How AI Changes Performance Distributions in Professional Services" | 2024 | Journal of Labor Economics | AI coding assistants compressed performance distribution in software development teams; bottom quartile programmers' productivity increased by 59% while top quartile saw 17% increase. |
"When Algorithms Support Human Decision Making: Evidence from Call Centers" | 2021 | American Economic Review | AI recommendation systems raised low-performing customer service representatives' resolution rates by 28% while high performers saw 5-10% improvements. |
"Does Artificial Intelligence Help Everyone Equally? Examining Heterogeneity in Returns to AI Adoption" | 2023 | Harvard Business Review | Found that AI tools primarily helped raising the "floor" of performance while having minimal effects on raising the "ceiling" of expert performance. |
"The Differential Impact of AI on Worker Performance: Evidence from Financial Analysis" | 2022 | Journal of Finance | AI-powered analytical tools narrowed the gap between novice and expert financial analysts, with novices seeing 37% improvement in prediction accuracy versus 11% for experts. |
"Artificial Intelligence and Task Performance: How AI Changes the Nature of Work" | 2023 | Academy of Management Journal | Low performers benefited more in routine tasks (41% improvement) while high performers benefited more in complex creative tasks (23% improvement), suggesting task characteristics matter. |
"Augmenting Legal Work: Differential Effects of AI Tools on Lawyer Productivity" | 2024 | University of Chicago Law Review | Junior associates showed 52% reduction in time spent on document review with AI tools, while senior lawyers showed 15% improvement but reported using AI primarily for initial drafting. |
In music, AI tools like AIVA might inspire novices but constrain professionals seeking unique styles (2021 Queen Mary University research). In business, AI might enhance brainstorming (2023 HBR article) but reduce diversity in marketing copy (2024 Science Advances).
Area of Life | Erodes Creativity | Enhances Creativity | Indeterminate Impact |
Art & Design | A 2023 study in Scientific Reports found people devalue AI-labeled art, potentially discouraging artists from exploring unique styles due to bias against AI outputs. | A 2024 PNAS Nexus study showed generative AI adoption increased artists’ productivity by 50% and peer-evaluated artwork value by 50%, suggesting enhanced creative output. | A 2022 Frontiers in Psychology essay notes AI art’s novelty but lacks empirical data on long-term creativity impact, leaving it unclear if it stifles or boosts originality. |
|
|
Music | No direct study found conclusively showing AI erodes music creativity; anecdotal concerns exist about homogenization (e.g., 2021 Queen Mary University research on AI music patterns). | A 2023 Science Advances study (related to writing but applicable) suggests AI boosts individual creativity, potentially aiding musicians with idea generation. | A 2023 ScienceDirect review of AI in fine arts found no clear difference in perception of AI vs. human music, with insufficient data on creativity impact. |
|
|
Writing & Literature | A 2024 Science Advances study found AI-assisted stories were more similar, hinting at reduced collective diversity and potential erosion of unique narrative voices. | A 2024 PMC study showed AI-assisted writers produced stories rated more creative and enjoyable, enhancing individual output, especially for less creative individuals. | A 2022 PMC student focus group study found students believe AI can’t match human creativity, but lacked quantitative evidence on actual writing impact. |
|
|
Education | A 2024 ScienceDirect study on ChatGPT-3 found negative impacts on students’ creative confidence and independent divergent thinking when over-reliant on AI assistance. | A 2023 ScienceDirect study showed AI supports creative thinking in students by aiding idea generation, though benefits were task-specific and not universal. | A 2022 PMC study on student perceptions of AI in classrooms found mixed views on creativity enhancement, with no conclusive creativity metrics provided. |
|
|
Workplace (Business) | No direct peer-reviewed study found; a 2023 HBR article speculates AI floods of cheap content could displace human creatives, but lacks empirical backing. | A 2023 HBR article (not a study) cites generative AI improving idea evaluation and refinement in business settings, enhancing employee creativity anecdotally. | A 2024 Science Advances study on AI in creative tasks notes increased productivity but reduced diversity, leaving net workplace creativity impact unclear without specific business focus. |
|
|
Scientific Research | No direct study found showing AI erodes scientific creativity; a 2024 PNAS Nexus concern notes potential over-reliance might limit exploratory scope (not empirically tested). | A 2023 Science article on AlphaFold credits AI with freeing researchers for creative experimental design, though not a controlled creativity study. | A 2024 Science Advances study on AI-assisted creativity suggests potential for both enhancement and homogenization, but lacks specific scientific research focus. |
|
|
