One of the most talked about ethical issues currently is in the context of the FIFA World Cup. It started with a controversial red card given to the U.S. soccer team’s star player Folarin Balogun, resulting in an automatic ban from playing against Belgium. A very human decision leading to the application of a long standing rule. Enters U.S. President Donald Trump, who intervened to reverse the decision and was granted his wish by FIFA President Giannu Infantino. That political interference comes as FIFA was already facing growing pressure to address its ethics records.
These were all very human decisions.
Algorithmic aversion
The FIFA World Cup backlash is only one recent example. Ethical issues are not just in sportlitics. They surface regularly in all sectors of the economy and society. The once behemoth U.S. energy, commodities, and services company Enron Corporation collapsed after relying on what turned out to be accounting fraud. Executives ended up sentenced to prison. Then premier global accounting and consulting firm Arthur Andersen collapsed in 2002 in the process. This is a well studied case of ethics and governance.
Still, people fundamentally trust humans more than algorithms for ethical advice. This is what Wharton School Professor Christian Terwiesch found in a recent research titled Advice quality and source disclosure shape trust in AI-generated ethical advice.
Our results reveal that before observing the advice, humans display a strong algorithm aversion in this context, with 72.6% of participants preferring to be advised ethically by humans.
Could this aversion impact judgement of the quality of advice and the willingness to accept it?
AI Ethical Guidance Perceived as Equivalent to Human Expert Advice
The experiment established that individuals perceive the quality of AI-generated ethical advice to be on par with that of a recognized human expert. By testing the perceptions of diverse groups—including laypeople, MBA students, and a specialized panel of scholars and clergy—researchers found that AI-produced guidance was deemed just as valuable as advice provided by Dr. Kwame Anthony Appiah, the renowned New York Times “Ethicist” columnist.
The study’s pilot results revealed that when participants were blinded to the source of the advice, 57% of the nearly 200 participants preferred the AI-generated responses over those written by the human expert. This preference was most significant among laypersons (59.6%), but was also shared by the majority of the ethics experts (55.5%) and MBA students (51.3%)
Are humans willing to accept AI ethical advice?
Liking the quality of AI advice is different from accepting the advice.
It turns out the algorithm aversion bias significantly decreases once people experience the actual quality of the AI’s responses.
After being exposed to the quality of the AI-generated advice, however, algorithm aversion is reduced substantially to 53.2%.
So we’ve gone from an aversion rate of 72.6% to 53.2%.
That percentage decreases to 46.3% when the source of advice is hidden. That is, if people are exposed to the advice without knowing it comes from AI.
Yes, exposure to AI increases acceptance of advice
Taken together, our findings suggest that while humans initially exhibit strong resistance to AI-generated ethical advice, this aversion significantly diminishes when they experience the quality of AI guidance firsthand, indicating that trust in algorithmic ethical reasoning may be more malleable than previously assumed and could evolve as people gain direct exposure to AI’s capabilities in moral decision-making
Three takeaways for financial advisors communications
1. Lead with reasoning
The findings suggest that trust in AI is “malleable” and depends heavily on the substance of the advice. For communications, this means organizations should consider presenting the rationale and reasoning behind a recommendation before disclosing its machine origin. Because Large Language Models (LLMs) use natural language to articulate reasoning, their advice appears more transparent and “human-like,” which helps soften initial skepticism.
2. Overcome ‘Algorithm Aversion’ through direct exposure
Financial advisors should realize that initial resistance to AI is high (nearly 73%), but it halves once clients actually experience the quality of the output.
This means that instead of merely telling clients you use AI, show them the expert-level outputs it generates. The findings indicate that “witnessing the expert-like quality” is the primary driver for reducing aversion and increasing acceptance.
3. Humanize AI Integration in Emotionally Sensitive Contexts
The research found that humans still have a preference for human experts in dilemmas that are “less emotionally loaded” or where the expert provides “clear directives” rather than abstract reasoning.
For communications, focus AI use on the “analytical power” and data-driven “expert” insights, but keep the human advisor at the center for final moral and personal judgments. The goal is to create a “hybrid” model where AI provides the high-quality ethical reasoning while the human provides the social legitimacy and consciousness that AI lacks.
Contact yperspective.ca to discuss how to efficiently communicate in the age of AI.
