Recently on The Science of Personality, Roberto Affonso Santos, owner and managing director at Ateliê RH, an authorized Hogan distributor in Brazil, shared his research comparing the personality characteristics of Brazilian managers to those of global managers. Roberto also asked ChatGPT 4.0 to describe Brazilian managers, then compared the AI-generated description to Hogan data. His findings from the study call into question both ChatGPT and human assumptions about leadership in Brazil. Let’s take a look.

Researching Leadership in Brazil

The origin of this study emerged from two converging points: Hogan’s 2023 global norm update and ChatGPT’s explosion of popularity.

First, Roberto decided to compare data about Brazilian managers to global data. “I found more divergence than convergence,” he observed. Then, Roberto asked ChatGPT to list the characteristics of Brazilian managers. “To my surprise, it gave a list of elements that are stereotypes about resilience that you hear from people in other countries,” he said.

These results reaffirm that ChatGPT is not a fact-finding machine. Its response about Brazilian managers is based on data contaminated with stereotypes about Brazil.

Personality Characteristics of Brazilian Managers: ChatGPT vs. Hogan Data

Large language models such as ChatGPT are not reliable sources of facts about the personality characteristics of Brazilian managers (or any other group, for that matter). Insights about personality should instead come from validated personality data.

ChatGPT: ‘Flexible and Adaptable’

ChatGPT described Brazilian managers as flexible and adaptable. The Brazilian economic, social, and political spheres can be unpredictable. Thus, the description matches a common perception that leadership in Brazil requires being adaptable to changing environments. “This was not confirmed by Hogan data,” Roberto said.

Hogan data described Brazilian managers as rule following, conservative, change resistant, risk averse, and less imaginative compared to global managers. It also indicated they prefer stability and security. How can these differences be?

When data do not match our expectations, we first evaluate whether our expectations are built on stereotypes. What assumptions or biases might influence what we expect? We can also explore how the data compares with other facts by investigating the contrasts that might affect our perceptions. “When we put that contrast on, we see things a little bit differently,” said Ryne Sherman, PhD.

In this case, ChatGPT is trained on a body of text from publicly available sources on the internet—sources that can (and do) include biased information. Its responses are based on that text. By contrast, the Hogan data about Brazilian managers are filtered against the personality data of other managers via Hogan’s global norm. The global norm is a set of personality assessment data that is representative of working adults around the globe, across industries, and across demographics. “With [Hogan’s] global norm, we are getting to the core of human differences,” Roberto pointed out.

ChatGPT: ‘Enthusiastic, Passionate, and Charismatic’

ChatGPT also described Brazilian managers as enthusiastic, passionate, and charismatic. That description didn’t align with Hogan data either. “Brazilian managers are often described by others as being extraverted, expressive, charming, persuasive, and using strong emotional commitment to engage others,” Roberto said.

Hogan data described Brazilian managers, compared to global managers, as less affiliative, less desirous of social interactions, less sociable, and less extraverted. Global managers also tended to be higher in charismatic traits that can create immediate interpersonal impact.

Not only does Brazil have a different official and national language than other South American nations, but it also has many regional cultural differences that can diverge from stereotypes. Generalizations about personality characteristics, such as those generated by ChatGPT, often stem from unfounded perceptions of “national character.” In fact, very little data exist to support these stereotypes.

Roberto gave an illustration to demonstrate the pitfalls of assuming that all Brazilians conform to the classic Latin American stereotype of passion and charm. A hiring manager at a multinational company interviewed a candidate for leadership in Brazil but initially rejected the candidate because they didn’t seem to have charisma.

“There’s an expectation that Brazilians should be expressive, great impact, great charisma,” Roberto said. “When that doesn’t show up, it frustrates people.” Roberto explained to the hiring manager that some Brazilians build their impact with time, not on the first impression. After a second interview, the Brazilian leader was hired—and they became a vice president in the company.

Other Personality Differences of Brazilian Managers

Compared to other managers globally, Brazilian managers’ scores differed on a few scales across Hogan’s three personality assessments: the Hogan Personality Inventory (HPI), Hogan Development Survey (HDS), and Motives, Values, Preferences Inventory (MVPI). Specifically, Brazilian managers had lower scores on MVPI Hedonism, higher scores on HPI Interpersonal Sensitivity, and higher scores on HDS Bold. (These characteristics also conflict with the ChatGPT description of Brazilian managers.)

• Lower Hedonism – Brazilian people may have the general reputation of being soccer, beach, and festival lovers, but Brazilian managers scored lower than their global colleagues on the Hedonism scale. Roberto said they seek to avoid risks and follow rules, valuing professional, formal work environments.
• Higher Interpersonal Sensitivity – Brazilian managers tend to have a relational and collaborative nature with good listening skills, empathy, and concern for others. This everyday strength can sometimes cause Brazilian managers to avoid confrontation, direct communication, or tough conversations. “We tend to sugarcoat an issue not to hurt others,” Roberto observed.
• Higher Bold – Despite tending to be less assertive in their day-to-day professional lives, Brazilian managers showed a higher score on the Bold scale of the HDS. The HDS describes derailers, which are behaviors that emerge during stress or complacency. In this case, a high Bold score can indicate that someone might seem entitled, arrogant, and overconfident when they stop self-monitoring.

Avoiding Bias in People Decisions

“ChatGPT tends to base its responses on common sense, popular publications, stereotypes,” Roberto said. If ChatGPT were to make a candidate selection decision based on its assumptions about Brazilian managers or leadership in Brazil, it would be susceptible to similar biases as a human interviewer.

Comparing AI-generated descriptions to Hogan data can be very revealing about the differences in popular perceptions and scientifically measurable reality. Errors in this regard can often lead people astray and create conflict, which is why accurate, reliable, and valid personality assessment data are essential.

Simply put, AI-generated information is no replacement for sound science. “I tend to rely much more on the objective, impartial, scientifically validated measures that we have,” Roberto said.

Listen to this conversation in full on episode 88 of The Science of Personality. Never miss an episode by following us anywhere you get podcasts. Cheers, everybody!