Faculty Research

Faculty Research Briefs

• Dangers of Speech Technology for Certain Employees, Dr. Min Geiger, Assistant Professor of Management
• Climate Risk Premium: Evidence from Commodity Options, Dr. Rui Liu, William and Helen Lyons Faculty Fellow in Finance; Associate Professor of Finance
• Digital Nudging and Donations, Dr. Georgiana Craciun, Associate Professor of Marketing
• AI Use and Consumer Well-Being, Dr. Christina Kuchmaner, Assistant Professor of Marketing
• Debt and Stock Market Participation: Evidence and Implications of Impulsivity, Dr. Charles Favreau, Associate Professor of Finance

Dangers of Speech Technology for Certain Employees

Dr. Min Geiger, Assistant Professor of Management

AI is now an integral part of many organizations, reshaping how people work and how organizations manage their employees. Although cutting-edge AI technology offers unprecedented efficiency and productivity, it may also damage organizations’ efforts to reduce barriers facing employees from marginalized groups.  

Research published in Nature Machine Intelligence by Dr. Min Geiger and her collaborators focuses on speech technology including systems like Google Translate and Microsoft’s Intelligent Speaker.  

The authors identified three significant issues that organizations may encounter when adopting speech technology. These issues can perpetuate existing barriers in the workplace and cause disproportionate harm to employees from marginalized groups.

Hidden Performance Barriers
Speech technology can introduce hidden performance barriers that increase existing disparities for employees from marginalized groups. These performance barriers are a byproduct of the technology’s limitations in accurately understanding and interpreting speech from marginalized speakers (e.g., nonnative speakers, employees with regional accents or dialects, employees with certain voice-related impairments), causing them to feel they are not heard in the workplace.     

Sensitive Biometric Data
Speech technology captures a large amount of sensitive biometric data, including individual characteristics that collectively form a voiceprint. Although some organizations collect biometric data for legitimate purposes (e.g., authenticating identity), others may misuse them by identifying underlying health conditions.

Inclusivity Gaps in Anthropomorphic and Interactive Design
Most speech technology designs and anthropomorphic features tend to overlook characteristics of people from marginalized groups (e.g., non-Westerners, nonnative English speakers). This oversight may promote social disconnection among employees from marginalized groups, leading them to feel they are socially excluded in the workplace.

Dr. Geiger’s research also provides practical guidance for organizations that seek to implement speech technology responsibly and strive to support employees from marginalized groups. For example, organizations should be aware of the potential performance barriers associated with speech technology. They should also be transparent about biometric data collection and proactively develop internal controls to ensure ethical use of the data. Finally, organizations should take time to truly understand how marginalized employees think and feel when they are interacting with speech technology.

Speech technology can disproportionately disadvantage marginalized employees in the workplace. It is important for organizations to address new forms of marginalization in the age of AI.
—Dr. Min Geiger

Climate Risk Premium: Evidence from Commodity Options

Dr. Rui Liu, William and Helen Lyons Faculty Fellow in Finance; Associate Professor of Finance

Understanding how risks are reflected in financial market prices has been an important area of research, particularly as risks related to climate change have attracted growing attention in recent years. Ongoing work by Dr. Rui Liu examines how climate risks are reflected in commodity market prices.

This study finds that climate risk premiums—the extra return investors demand for climate-related risk—increase with low-to-moderate levels of uncertainty. However, they decline as uncertainty becomes more extreme, creating an inverted U-shaped relationship. This means that premiums rise at first, then fall as uncertainty increases further. In the short term, changes in policy uncertainty are the main factor driving these premiums, with sustained uncertainty increasing how sensitive prices are to that risk. Market participants like investors and traders responded more strongly when uncertainty increased than when it decreased by a similar amount.

By demonstrating that climate policy uncertainty affects risk premiums in uneven and nonlinear ways, Liu’s findings suggest that clearer and more predictable policy guidelines could help stabilize market expectations. Consistent environmental regulations are vital for reducing uncertainty, helping markets price climate risk more accurately, and ultimately encouraging more investment in sustainable assets.

By integrating climate factors into market practices, opportunities for sustainable investment opportunities expand. As regulators increasingly emphasize climate disclosures and risk management, quantifying climate risk premiums can inform the development of market practices and regulatory frameworks that align finance with a sustainable, low-carbon future.
—Dr. Rui Liu

Digital Nudging and Donations

Dr. Georgiana Craciun, Associate Professor of Marketing

Understanding how users interact with online donation forms is important for non-profit organizations seeking external funding. Charitable contributions can be positively or negatively affected depending on how information is presented.

Nudges are a common design element in online donation forms. Two common types are default and friction nudges. A default nudge is a preselected option that occurs without user input. For example, when a $50 donation amount is automatically selected in a form, or the monthly recurring donation box is already checked. Default nudges rely on convenience and perceived recommendations to emphasize specific donation options. A friction nudge works the opposite way by adding extra steps to discourage a particular choice. A friction nudge may require users to type in a custom amount for a smaller donation or open a separate menu, for example.

Research published by Dr. Georgiana Craciun studies how users respond to these default and friction nudges, as well as transparency about nudge methodology.

Craciun’s findings show that friction nudges lead to significantly higher donation amounts, while default nudge returns are more marginal. Combining default and friction nudges sees significantly diminished returns, and nudge transparency has a negligible effect. Transparency does not meaningfully reduce donations and alleviates concerns about donor backlash.

Overall, these findings suggest that non-profit organizations seeking donations may see more success by using friction nudges under certain conditions.

Offering a higher amount upfront and placing smaller amounts behind an extra step increases giving. Explaining this design openly preserves trust while keeping choice intact.
—Dr. Georgiana Craciun 

AI Use and Consumer Well-Being

Dr. Christina Kuchmaner, Assistant Professor of Marketing

Some people are threatened by AI because they believe it may eventually exceed human intelligence and capabilities, making human input obsolete. Research published by Dr. Christina Kuchmaner examines factors that decrease user’s AI threat perception.

When using AI, consumers automatically apply social rules and expectations to machines, even after one interaction. These humanlike interactions can quickly lead to a feeling of connection with their AI tool as a humanlike partner. Kuchmaner’s research establishes the concept of “perceived AI relatedness,” which is a user’s generalized sense of connectedness to AI. Her research shows that AI relatedness can reduce perceived AI threats through ingroup favoritism, which emerges because users come to view AI as an “us” rather than a competing “them.” Notably, however, the reduction does not occur if users feel forced to use AI.

The findings offer important practical and societal implications by demonstrating that fostering a sense of connection between consumers and AI can reduce perceived identity threats and resistance, ultimately improving adoption, satisfaction, and performance across the customer journey. Firms can leverage this by designing AI experiences that emphasize collaboration and personalization while also providing opportunities for voluntary engagement and human support when needed. At the same time, increased AI integration raises broader societal concerns, as reduced threat perceptions that may coexist with risks to consumer well-being. This could include emotional overreliance or psychological harm among vulnerable users. Accordingly, these insights highlight the need for responsible AI design, consumer awareness, and policy safeguards to balance the benefits of AI adoption with protections against its unintended consequences.

We’re at a point where AI can feel less threatening and more relatable. But for younger users especially, we need to ensure these technologies support well-being without replacing meaningful human connection.
—Dr. Christina Kuchmaner 

Debt and Stock Market Participation: Evidence and Implications of Impulsivity

Dr. Charles Favreau, Associate Professor of Finance

Economic theory suggests nearly all individuals should have some investment in the stock market, but only about half of U.S. adults do. Participation rates are also declining among younger adults. Several explanations have been offered for low stock market participation, including participation costs, financial literacy, wealth, and social factors.

Many studies rely on people’s assets or wealth (assets minus debt) when examining stock 
participation. However, the former implies that debt has no impact on participation, while the latter assumes that a $1 increase in assets and a $1 decrease in debt equally affect participation. Research published by Dr. Charles Favreau finds overwhelming evidence that debt negatively influences stock market participation nearly twice as much as the positive impact of assets. Examined further, the study finds that impulsivity explains part of the relationship between debt and participation. Other financial research has established a strong connection between impulsivity and high levels of debt.

Favreau’s research could help policymakers understand and address the widening wealth gap. It also indicates that future stock market participation studies should attempt to discover more precisely why having debt seems to have a negative connection to stock market participation.