Artificial Intelligence
Generative AI May Reduce Corporate Herd Behavior
AI Is Moving From Productivity Tool to Capital Allocation Tool
While retail focus remains heavily locked on AI’s impact on worker productivity, institutional research is turning to its systemic impact on capital allocation and market reflexivity.
Generative AI is expected to boost companies’ productivity by automating a lot of tasks that were so far dependent on human labor: customer service, legal advice, translation, proofreading, data entry, quality insurance, etc.
The effect this change will have on job markets and the competitive landscape has been the main focus of investors and the public about the arrival of generative AI.
But it is likely that the impact of AI will go much further. For example, human top performers are more likely to embrace new technologies and to put them to innovative use. In general, it is increasingly considered that AI is a “force multiplier” in the knowledge economy.
So it stands to reason that generative AI could have a major impact on decision-making at the highest level of large corporations, where management will use it to support or influence their decisions.
A recent article by researchers at the Capital University of Economics and Business in Beijing (China) seems to confirm it. The article claims that generative AI may reduce corporate herd behavior by enhancing management’s decision-making independence.
The article was published in International Review of Economics & Finance1 under the title “Blind imitation or rational decision-making? The impact of generative artificial intelligence on corporate investment herd behavior“.
What Corporate Investment Herd Behavior Means
It is a well-studied phenomenon that corporations can behave as a group, each individual company making decisions very similar to its peers during a given period. This is notably a strong effect in firms’ investment decisions, where imitating competitors tends to be done for a multiple of reasons: keeping away key technologies from competitors, avoiding the risk of looking passive in changing market conditions, reacting to innovation, etc.
A key driver of this attitude is that decision-making at different corporations is often driven by shared information.
“When firms rely on highly similar information sources or place greater weight on public signals, even individually rational decisions may generate systematic distortions at the aggregate level.”
But this corporate herd behavior can also lead to massive mistakes. A promising idea can turn into a devastating bubble if pushed too far. A new technology might prove not as useful as hoped for. A new management method might bring unforeseen new problems.
So while it is probably impossible to fully avoid, herd behavior is overall not a good thing, and anything that reduces it will likely be a benefit for corporations and the economy as a whole.
What the Study Found
The study focused on investment decisions by corporate entities, using Chinese firms as the data sample.
What this study discovered is that, down to the firm level, there is evidence that the adoption of generative AI affects investment herding.
It not only reduces herd behavior in corporate investment decisions but also enhances decision-making independence in general.
This would indicate that adoption of generative AI can have a widespread effect on corporate strategy, much beyond just improving productivity at the worker and department levels.
It also found that the company’s ownership and industry matter. The effect was stronger in non-state-owned enterprises and high-tech industries, while it was weaker in non-high-tech industries and state-owned enterprises.
Why Generative AI Could Reduce Blind Imitation
It has been previously expected that the deployment of AI could radically reshape the process and roles of management in corporations.
“Because artificial intelligence emphasizes model-driven and algorithm-assisted decision making, its use may redefine managers’ roles in the decision process by shifting decision making toward systematic analysis rather than individual experience and judgment.”
When it comes to this study, an important context to understand is that investment decisions are typically made in environments characterized by incomplete information.
A firm must assess not only the projects’ cash flow prospects, but also anticipate changes in market demand, competitors’ actions, customers’ preferences, and macroeconomic conditions. For each of these data points, information available is more often than not incomplete or contradictory.
To add further complexity, firms differ substantially in their ability to acquire and process this information.
This complexity is often believed to be why observation of peers plays such a big role in the final, real-world decision process.
“Peer firms’ investment behavior becomes an observable and summarized external signal. By observing whether peers invest, how much they invest, and the timing of their investments, a firm can partially infer market-wide expectations and reduce the uncertainty surrounding its own decision making.”
In financial markets, this herd behavior has been described as reflexivity, a self-reinforcing feedback loop feeding on imitation and shared perception, and can be a key reason for the emergence of financial bubbles.
Adoption of generative AI pushes for more unique decisions by reshaping firms’ information-processing practices. As the information pool from which the decisions are taken is richer, the decisions taken are more diverse than before.
“Algorithm-driven analysis enables continuous information updating, allowing firms to form relatively clear assessments at earlier stages and reducing the incentive to wait for others to act in order to learn.”
Why AI Could Also Create New Forms of Herding
So it seems that AI can reduce the dependency on imitating peers, competitors, and industry trends by providing more original and actionable data.
Although this should not be overstated, as it is possible that AI adoption correlates with hidden, firm-specific traits that explain the difference in investment decisions. For example, a more innovative company with a more insightful management would likely to both take more original investment decisions and lead in AI adoptions.
“Factors such as a firm’s management quality, technological innovation capabilities, or market positioning may affect both its adoption of generative AI and its investment behavior. If these factors are not included in the model, it could result in either an overestimation or underestimation of the effect of generative AI.”
It should also be noted that the conclusion is drawn from observing Chinese firms only, and different cultures and business ecosystems might have different results.
The result of this study also does not mean that AI adoption could not lead to new forms of herd behavior.
For example, using similar AI models might lead to similar conclusions across firms.
Another potential issue is that AI data are not generated in a vacuum. If industry-wide data points toward some trends, it is likely that even different AI models will pick up on it, each independently, and lead to the same conclusion given to the firm’s management. So AIs themselves could prove as sensitive to herd behaviors as a human manager alone.
What This Means for Investors
For investors, the discovery that AI might break or alter investment herd behavior led to multiple conclusions.
The first one is that early adoption of AI, not just for menial tasks, but for intelligence and management decisions as well, could prove a decisive edge. So they might want to invest in priority in such companies.
It could also mean that this might be a serious blow to investment strategies relying on momentum and herd behaviors. If AIs make everyone’s decisions more unique, we might see less money rushing at once in a sector-wide boom, or specific technology or investing themes suddenly becoming “the new hot trend”.
Conclusion: Better Decisions Depend on Better Governance
So it seems that the adoption of AI for guiding investment decisions is reducing herd behavior. This was especially prevalent for privately owned tech firms, and further investigation would be needed to clarify why.
It is also possible that early AI adoption is a reflection of just better and more flexible governance, and not so much a direct cause of more original investment choices.
Unsurprisingly, it is clear that better governance leads to better decisions, be it strategic moves, investment decisions, or smartly leveraging AI data for more original choices.
This emphasizes that for investors in the stock market, studying a company’s fundamentals is not just about analyzing financials, market shares, or innovation. It also includes a solid assessment of governance and the quality of management’s decisions.
So, truly independent thinking, often proven by unique and original investments, is a very important criterion, be it triggered by openness to use AI or independent human-only assessment.
Study Referenced
1. Yongxiang Wang, Zhanhong Wu, Jun Zhang. Blind imitation or rational decision-making? The impact of generative artificial intelligence on corporate investment herd behavior. International Review of Economics & Finance. 19 March 2026. 10.1016/j.iref.2026.105139.
