Will AI replace marketing managers? Q&A with Professor Zsolt Katona

A photo shows a man in purple dress shirt behind computer monitors
Professor Zsolt Katona

Even though AI assistants such as Apple’s Siri and Amazon’s Alexa have been in widespread use for years, the release of ChatGPT in late 2022 was astonishing. The bot’s advanced language understanding and human-like responses opened the public’s eyes to the rapid pace of AI development, set off a wave of excitement and apprehension, and ramped up competition among tech giants and upstarts to deploy new AI technologies. Apple joined the race this week with the announcement of Apple Intelligence, which will put a powerful AI chatbot into the pockets of hundreds of millions of iPhone users worldwide.

Professor Zsolt Katona, who holds PhDs in both computer science and marketing, began using generative AI in 2019 to write scripts for his Berkeley Executive Education course. That year, Katona also developed and began teaching the Business of AI course for MBA students. He has recently focused his marketing research on AI as well.

We talked with Katona, the Cheryl and Christian Valentine Professor, about misconceptions, business applications, and how AI is influencing marketing.

Berkeley Haas: Right now, what do you think are the biggest misconceptions about AI?

Zsolt Katona: I guess one of the biggest misconceptions is just the word ‘generative’ because people are applying it to everything—that’s the hype—and definitions aren’t clear.

“…One of the biggest misconceptions is just the word ‘generative’ because people are applying it to everything—that’s the hype—and definitions aren’t clear.”

Tell me more, because I could be one of those people.

People use generative to mean that the application generates something. But many of the uses people are familiar with are more like a kind of search. And in terms of beliefs about how valuable different applications are, the family of generative AI is overestimated. Most of the applications that are the most lucrative are not generative in nature.

I read an article that said Mastercard uses generative AI for fraud detection. But it turns out that they use a transformer model—these are (neural networks) behind language models. Transformers were originally designed for generative applications, but they have nongenerative uses, and Mastercard’s use was nongenerative. It really is just a tool that detects outliers, suspicious transactions.

Other than confusion about terminology, the biggest misconception is beliefs about the ability of these things to work fully autonomously. That’s essentially nonexistent in most applications. What you have to do for pretty much every application is figure out how to make the AI portion and the humans work together.

I’ve heard it said that we’re still in the hype phase, and many companies are still trying to figure out lucrative commercial applications.

They’re figuring it out, and it’s just a matter of time. Some of the fancy stuff is not there yet because companies are having problems with getting their data infrastructure ready. Their data is messy, it’s not in a format that allows them to easily use simple AI applications—or they might not even own the data. But it’s the nonflashy stuff that’s most lucrative. For example, using cameras in a factory to detect manufacturing defects. There’s nothing generative AI about it—it’s just looking at little differences in those pictures.

Illustration shows an hands typing on a laptop with a digital image with computer code and the letters AI
Image: AdobeStock

What’s an example of something businesses are doing well, or a significant change, in marketing?

That depends on whether you mean something that’s widely used versus something that’s flashy and innovative. I like the ones that are useful, such as all the personalization that’s happening on a large scale. For example, customized videos for each product where somebody explains exactly what the product does. It’s happening on Chinese ecommerce sites, but it’s coming to Amazon very soon. It’s just impossible to do something very personalized to the consumer without this kind of technology.

In your recent research, you looked at whether market researchers can replace human participants with ‘synthetic respondents’ generated through a large language model (LLM). What did you find?

We only tried a couple of product categories, but it worked pretty well. We had just three variables—age, gender, and income. There was 75% to 90% agreement with human data.

Are we almost at the point where market researchers can accurately use AI for product research?

They’re already doing it. Our paper was about validating that this works, and for that, you still need human data to compare. The promise of these ‘synthetic respondents’ is that you can get very specific types of responses that would be otherwise hard to get from humans, and at a much lower cost. Let’s say, a person who earns a million-plus dollars and lives in a fully electrified home and drives a cyber truck. You can ask the AI to pretend to be that person and answer questions about perceptions about cars. You’ll get a response, but it’s still hard to validate because you have to find that human to compare it to.

“The promise of these ‘synthetic respondents’ is that you can get very specific types of responses that would be otherwise hard to get from humans, and at a much lower cost. Let’s say, a person who earns a million-plus dollars and lives in a fully electrified home and drives a cyber truck.”

So the more specific you get, the less accurate it might be because it would be hard to create a sample of people like that.

The question becomes, how bad it would get? Would it be totally random, or would it give you some idea?

How would this work with a new product—such as a brand-new car brand with no information from humans on the internet, or a really innovative product that doesn’t exist yet?

In theory, if you can accurately describe the new brand or new product, the language model could make those inferences. For example, you can tell AI to create an image of a cat with a mask on, right? You don’t need examples of cats with masks on. You need examples of cats, and examples of masks, and examples of some kind of human or animal with a mask on so that it understands these three things.

How does AI do in coming up with new products?

We actually have a working paper on creativity for product ideas with humans versus AI. We do find that AI is better, but if you look only at the top answers, the difference is much smaller. And again, it’s more of a human data problem because, with the AI, we can get the best models to generate ideas, but we can’t ensure the most creative humans are going to be in our pool of human participants. It’s still very likely that if you somehow managed to get all the humans in the world to participate in our study, the best ones would be better than AI.

You teach the business of AI to MBA students and through Berkeley Executive Education. With things moving so fast, what skills do managers working with AI technology need?

They need to understand the fundamentals of how it works. That’s the No. 1 thing. It’s even better if they have some coding skills, and I do make them go through an exercise with code, so they have at least a feeling for what it looks like and for the building blocks. Obviously, anyone who studies engineering will understand it in great detail, but for nontechnical people, just understanding how it works helps them a lot. Other than that, they need to understand how to manage technology, which is not that specific to AI. If you put together those skills with some understanding of how the specific technology works, it’s tremendously helpful.

Can nontechnical people learn enough to be effective?

My marketing colleague who teaches marketing analytics likes to say that it’s easier to teach managers analytics than to teach data scientists to be good managers. I share that thought, and again, they don’t have to be as technically advanced as the engineers. But they should understand how the data goes in and how it results in a desired outcome.

My advice is that managers should learn enough about how it works to talk to the people who make these things, especially with respect to the data needs. What I call the “objective function” of the model is what it should do. That’s just an equation, but translating that equation to the business objective is a critical task that somebody has to do, and it’s not going to be a data scientist. It’s rarely going to be the engineer.

Is that pretty much the same as for technology management in general?

Well, the difference is in how AI works, and specifically that it learns from examples. So you have to think hard about what those examples are, and you have to think hard about how you train it. How should the so-called error or loss function be specified? What are you aiming for, and when do you say it’s good enough?

Will there be jobs for marketing managers without engineering backgrounds?

I think there will be. Marketing is such a subjective topic that it’s hard to evaluate all the things AI needs to do. It comes down to a lot of human judgment. If AI can do every job in the world, then yes, marketing people will be replaced as well. But it’s very hard to show that a machine can do the work better than a human.

“Marketing is such a subjective topic that it’s hard to evaluate all the things AI needs to do. It comes down to a lot of human judgment.”

Because of all the different aspects of it?

Yes, it’s a very complex type of work. And then, because of the complexity, you need a lot of people who know how to use these tools. So, their jobs might change, but they will have jobs for sure. Everybody was saying a few years ago how the blue-collar jobs would be replaced, and now they are instead talking about all the white-collar jobs. But neither is happening, really. Some tasks are being replaced, yes, but people will still have jobs—although they will be different.

Friends with health benefits: How the buddy system pays off when pursuing goals

Two women wearing leggings sit on the floor of a gym stretching.
Image: AdobeStock

Weekly targets, annual resolutions, five-year plans—all of them so troublingly elusive. With best intentions, most of us fail to stick with the goals we set.

Next time, consider pursuing them with a friend.

New field research by Assistant Professor Rachel Gershon, published in Management Science, suggests that pursuing our goals with friends may make them more attainable. Gershon, along with Cynthia Cryder of Washington University and Katy Milkman of the University of Pennsylvania, specifically looked at gym attendance and found that going with a friend—even with the hurdles of coordinating two schedules—increased visits by 35%.

“Despite adding the friction of working with another person, we saw people becoming more motivated and more likely to go,” Gershon says. “This illuminates how social incentives, which aren’t always taken into consideration, can help people overcome other barriers that stand in their way.”

The experiment recruited two groups of participants for a “Gym Bonus Month,” which lasted four weeks, from February 1 to February 28. Both groups paired up with a friend and were offered a $1 Amazon gift card for each visit to the gym. One group received this bonus every time they went to the gym, regardless of their friend’s activity; the other group only received the dollar if the two of them went together.

As noted, those who received payment only when they visited the gym with their friends doubled how often they went together, and increased their overall gym visits by 35%. Gershon and her colleagues concluded that the logistical costs of coordinating with someone else were eclipsed by two benefits. First, people enjoyed their visits more when the event was social, which made future visits more likely. Second, they felt a greater sense of accountability when meeting their friend at the gym.

“Our study identifies two types of accountability,” Gershon says. “People feel responsible to their friends, as they wanted them to get the reward, but they may also have reputational concerns that their friends would think less of them if they didn’t follow through.”

Social benefits

Although this might seem intuitive, when Gershon and her colleagues surveyed people about which of the two conditions they would prefer to be part of, the majority—more than 80%—said they would rather not have to coordinate their visits with a friend. While unsurprising in some ways, Gershon says, this suggests that people might readily see the drawbacks of coordinated visits but not recognize the potential benefits, from increasing motivation to creating stronger social bonds.

The researchers also found evidence that, when looking across both partners in a pair, this social attendance of the gym seemed to provide the greatest benefit for those who exercised less. Specifically, among the two friends, the one who exercised more frequently prior to the study saw a bump in how often he or she visited the gym. But the partner who exercised less frequently prior to the study saw an even larger bump in visits, suggesting these kinds of social incentives may be especially effective for distinct groups of people.

Beyond the context of this experiment, the findings illustrate how building a social dimension into desired behaviors can promote follow-through. Companies that want to increase employee engagement with skills training, for instance, might consider using a joint-incentive program. This could boost participation while simultaneously fortifying interpersonal bonds in the workplace.

The findings also present implications for another area that Gershon studies: referrals. Many places offer a free month of membership or some other incentive if you recruit a friend. “There are all sorts of contexts where people are trying to start a new hobby, a new exercise routine, and companies can encourage them through social networks,” she says. “This work shows that referrals may be a way for companies to not only engage additional customers, but to also increase the motivation of current customers.”

The paper:

Friends with Health Benefits: A Field Experiment
Rachel Gershon, Cynthia Cryder, and Katherine L. Milkman
Management Science, April 2024

How instincts lead us astray in deciding how to boost our chance of success

Image: AdobeStock

Say you’ve got an important presentation on Monday that could make the difference in whether a project gets green-lit or not. Would you spend a few extra hours prepping for it over the weekend to make sure it goes well, or blow it off and go out with friends? Now say the presentation has a low chance for success anyway—would that change your willingness to put in the hours?

If you’re like most people, it would, according to new research by Haas Associate Professor Ellen Evers, doctoral student William Ryan, and Stephen Baum, PhD 23, of the Olin School of Business (Ryan and Baum served as co-lead authors). They found that in all kinds of different situations, people underinvest time and resources to improve their chances when success is already unlikely, and yet overinvest in situations where that are likely going to be successful anyway.

“We make these improvement decisions on a daily basis,” Evers says. “Passing or failing a test, making sure we get a job, making a new friend. When you think about it, they are everywhere.”

Weighing chances of success

On a more consequential level, a doctor might have to decide how much time to invest in saving a dying patient, or a government official could face a decision on how much money to invest in preventing a terrorist attack. “There’s always a trade off between how much we’re willing to invest to maximize our chances for success.”

On a strict percentage basis, the initial chances for success shouldn’t factor into one’s decision on investing extra resources. Whether increasing from a 10 to 20% chance of success or an 80 to 90% chance, you still improve your chances by 10%. The researchers chalk up the disconnect to an emotional response: the fear of future regret. “If you already have a small chance of succeeding and you fail, then it’s easy to say, it wouldn’t have mattered anyway,” Evers says. “But if you had a 95% chance and you don’t make it, you’re going to be like, wow, it’s really my fault.”

The results point to the importance of taking the power of emotions into account when considering a decision, and—counterintuitively perhaps—putting emotions aside the most when stakes are highest.

In the first set of experiments in the paper, forthcoming in Psychological Science, the researchers asked people in the lab and online how much they would pay to increase their chances in a lottery of winning $10 by 10%. In all circumstances, the correct answer should be $1. And yet, they found systematically that people would pay as little as 50 cents when their initial chances were only 10%, and as much as $3 when starting with 80%.

Focus on the gain

Interestingly, the researchers could make this effect go away entirely by framing the situation to focus on the gain rather than the loss. After all, if you start with 10% and increase to 20%, you are doubling your chances to win; whereas if you increase from 80% to 90% you are only increasing your chance to win by 1/8. “It shows that people naturally focus on how much of the loss is their fault,” Evers says.

On the other hand, they could increase people’s willingness to overinvest in these decisions by ratcheting up the emotional intensity. In another experiment, they presented people with the option to buy a pill that could reduce their risk of getting a seasonal cold, once again offering situations in which the initial likelihood of getting the cold was high or low. Once again, they found that people were more willing to pay for the medicine when chances of getting the cold were already low—around $15 compared to $10.

But they also found people were willing to pay a premium when symptoms of the theoretical cold were more intense. Again, the researchers say, the finding emphasizes the power of regret, since if you get sick for a week, the consequences are higher than if you get sick for a day, causing people to invest more heavily. While that might make sense from an emotional standpoint, Ryan notes, it means people are more likely to make mistakes in valuing probability when the consequences are highest. “Sometimes we think people make mistakes because they don’t care enough and are not paying attention,” he says. “But this is just the opposite of that—they actually care too much.”

Experts do no better

Nor are people better at making these decisions when they have more expertise. In another experiment they asked actual doctors whether they would be more willing to invest time to improve a patient’s chance for survival from 10 to 20% or from 80 to 89%. Even though the first scenario was objectively better—representing a 10% increase compared to 9%—they found that more than half the doctors chose the second. Even when the researchers skewed the scenario more heavily, 45% of doctors still preferred improving the 2nd patient’s chances of survival by just 5% (from 80% to 85%) instead of improving the first patient’s chances of survival by 10%.

“It’s really important that doctors as a group are good at making these kinds of decisions,” says Ryan, the lead author on the paper. “If every doctor in the healthcare system did this, then fewer patients would survive overall.” At the same time, they speculate, the emotional costs of losing a patient are so high, doctors repeatedly learn the wrong lesson. “If they have a 10% chance of survival and pass away anyway, you never know if helping them more would have mattered,” Evers says. “But if a patient had a high chance of survival and they pass away, you are going to carry that for the rest of your life.”

One way of getting around that kind of distorted thinking is to think of these decisions in the aggregate—for example, investing in extra effort to save 10 extra patients for every 100, rather than increasing one patient’s chances for survival by 10%. Another strategy is to try and remove emotions from the situation as much as you can to minimize distortions. “If you want people to make better investments, then the worst thing you can do is say, ‘This is really important,’” Evers says. Since we all win some and lose some, reminding ourselves that it’s not all our fault can help lessen feelings of regret when things don’t work out—and help us achieve better rates of success over time.

The paper:

By William H. Ryan, Stephen M. Baum, and Elle R. K. Evers
Psychological Science, 2024

The Bigger Picture

Online images promote gender bias

Ten images of the same man dressed in clothing representing different careers. Clockwise from top left the careers are: traveler, gardener, chef, business executive, doctor, spy, painter, plumber, construction worker, and lounger. Next to each images are objects indicative of that role.These days, we’re bombarded with images on picture-packed news sites and social media platforms. And much of that visual content, according to new research, is reinforcing powerful gender stereotypes.

Through a series of experiments and with the help of large-language models, Assistant Professors Douglas Guilbeault and Solène Delecourt found that Google Images exhibit significantly stronger gender bias for both female- and male-typed categories than text from Google News. What’s more, while the text is slightly more focused on men than women, this bias is over four times stronger in images.

Delecourt says that most of the previous research about bias online has focused on text. “But we now have Google Images, TikTok, YouTube, Instagram—all kinds of content based on modalities besides text,” she says. “Our research suggests that the extent of bias online is much more widespread than previously shown.”

To zero in on gender bias in online images, Guilbeault, Delecourt, and colleagues designed a novel series of techniques to compare bias in images versus text and to investigate its psychological impact in both mediums.

First, the researchers pulled 3,495 social categories—which included occupations like “doctor” and “carpenter” as well as social roles like “friend” and “neighbor”—from Wordnet, a large database of related words and concepts. To calculate the gender balance within each category of images, the researchers retrieved the top hundred Google images corresponding to each category and recruited people to classify each human face by perceived gender.

Large-language models measured gender bias in online text by noting the frequency of each social category’s occurrence alongside references to gender in Google News text. The researchers’ analysis revealed that gender associations were more extreme among the images than within text. There were also far more images focused on men than women.

In another study, 450 participants searched Google for apt descriptions of occupations relating to science, technology, and the arts. One group used Google News to upload textual descriptions; another group used Google Images to upload pictures of occupations. Compared to those in the text and control conditions, the participants who worked with the images displayed much stronger gender bias associating women with arts and men with science (a bias linked to systemic inequalities in academia and industry)—even three days later.

“This isn’t only about the frequency of gender bias online,” says Guilbeault, the paper’s lead author. “There’s something very sticky, very potent about images’ representation of people that text just doesn’t have.”

Gym Buddies

Maximizing gym memberships

Two fit, young female friends laughing together after a gym workout. Photo: Flamingo Images/Adobe Stock
Photo: Flamingo Images/Adobe Stock.

Chances are, that gym membership you signed up for with the best of intentions on January 1 might already be underused. Next time, consider signing up with a friend.

New research by Asst. Prof. Rachel Gershon suggests that pursuing our goals with friends may make them more attainable. Gershon and colleagues from Washington University and the University of Pennsylvania specifically looked at gym visits and found that going with a friend—even with the hurdles of coordinating two schedules—increased visits by 35%.

In the experiment, participants were paired up with a friend and given either one dollar every time they went to the gym, regardless of their friend’s activity, or one dollar if the two of them went together.

The researchers concluded that two benefits eclipsed the logistical costs of coordinating with a friend. First, people enjoyed their visits more when the event was social, making future visits more likely. Second, they felt a greater sense of accountability.

“Our study identifies two types of accountability,” Gershon says. “People feel responsible to their friends, as they wanted them to get the reward, but they may also have reputational concerns that their friends would think less of them if they didn’t follow through.”

Beyond this experiment, the findings illustrate how building a social dimension into desired behaviors can promote follow-through. Companies wanting to increase employee engagement with skills training, for instance, could try a joint-incentive program to boost participation.

U.S. News ranks Berkeley Haas FTMBA Program #7 in 2024

The Berkeley Haas Full-Time MBA Program claimed the #7 spot among full-time programs in the 2024 U.S. News & World Report Best Business Schools ranking.

The FTMBA program moved up four slots to tie for #7 with the Yale School of Management and NYU’s Stern School of Business. Except for 2021 and 2023, the FTMBA has ranked #7 since 2019.

Meanwhile, the Evening & Weekend Berkeley MBA Program ranked #2 this year among part-time MBA programs. The Berkeley Haas MBA for Executives Program placed #7 among EMBA programs and is now the top executive MBA program at a public university in the nation. This ranking is based solely on ratings by business school deans and directors. 

The 2024 FTMBA ranking, released today, reflects positive changes that U.S. News made to its rankings methodology, said Haas Dean Ann Harrison. 

The ranking reflects all of the work Haas is doing to strengthen its programs and reputation, she said. “There are many different ways of evaluating a school, and rankings go up and down for all of us,” she said. “The change in the U.S. News methodology, with less emphasis on starting salary upon graduation, is a positive step.”

A few details on the rankings methodology used this year:

  • Employment rates at graduation – 7% weighted  (previously 10%)
  • Employment rates three months after graduation – 13% (previously 20%)
  • Mean starting salary and bonus – 20%
  • Ranking salaries by profession – 10%
  • Peer assessment score – 12.5%

Haas ranked #5 in salaries, which were ranked this year by profession (tied with Chicago Booth). Harrison noted that alumni accept jobs in a variety of industries, which logically means a variety of pay scales. 

“This is true for Haas, as well, where graduates prioritize where they can make the biggest impact, whether that is in consulting, product management, fintech, or by founding a new company,” she said. “I applaud U.S. News for taking into account the reality of the wealth of opportunities for a b-school graduate and comparing apples to apples across all the schools it surveys.”

Assessment by the school’s FTMBA peers was strong this year, at #7 (tied with Columbia) and the school ranked #9 for its recruiter assessment. Haas also had the highest GMAT score, tied at #1 with Stanford, Harvard, Wharton, Kellogg, and Columbia.

In specialty rankings, based solely on peer assessments, U.S. News ranked the full-time MBA program:

  • #4 in nonprofit
  • #4 in entrepreneurship
  • #4 in real estate
  • #7 in business analytics
  • #7 in management
  • #8 in finance
  • #10 in marketing

Q&A: Teaching the business of Taylor Swift at Berkeley Haas

young woman with long curly dark hair
Miaad Madeline Bushala, BS 25, co-teaches a DeCal on Taylor Swift.

Miaad Madeline Bushala, BS 25, likes Taylor Swift’s music but doesn’t consider herself a die-hard “Swiftie.” What’s more intriguing to her is Taylor Swift’s evolution as a business leader who continues to top the music industry.

Bushala is now tapping into how the 14-time Grammy winner built her fortune, co-teaching a DeCal at Berkeley Haas called “Artistry & Entrepreneurship: Taylor’s Version” with Sofia Mei Lendahl, a sophomore Data Science and Statistics double major. The pair were in their fourth week of teaching the 13-week class when Bushala talked to Haas News.

You came to this class with both a musical and a business background.

Indeed, I did. I was a vocalist in the Popular Music Conservatory at the Orange County School of the Arts (OCSA) alongside my brother who is a fantastic drummer and my biggest musical inspiration. I attended Grammy Camp twice for vocal performance, a camp where high school students across the nation learn from and collaborate with music professionals.

My business background comes from watching and helping my parents with their real estate business, and then of course all that I’ve learned since being a student at Haas.

What interested you most about Taylor from a business perspective?

I heard somebody say that “nothing about Taylor Swift is an accident,” and I truly do believe that. Particularly as a business student, Taylor’s story has been so fascinating to me. At the end of the day, her songs, albums, merchandise, tours, etc. are all products, and for a product to have a life of almost 20 years not only says something about Taylor’s brilliance as an artist, but as a  businesswoman. With that, I am interested in unraveling all those pieces about her and seeing what made her the success that she’s become.

I heard somebody say that “nothing about Taylor Swift is an accident,” and I truly do believe that.

How did you meet Crystal Haryanto, BA 23 (Economics, Cognitive Science, & Public Policy), who founded this class?

Crystal and I met through Lizzie Coyle, director of Major Gifts at Haas. Lizzie shared the excitement of the Taylor Swift course in the business school and I was encouraged to consider joining the team as the team was also seeking a business perspective. I was supposed to study abroad this semester in Spain, but this was my sign to stay and do something that I’d never done before.

As a business student, how did you help shape the class syllabus?

Taylor Swift performing
Singer Taylor Swift (AP Photo/Nati Harnik)

I asked the hard questions—for every concept in our syllabus, I ensured that there was a viable link to business. We wanted students to view Taylor as an entrepreneur who differentiates herself within a market, manages customer acquisition and sustains customer loyalty, and impacts multiple economies. We wanted them to think about how, as future entrepreneurs and business leaders, to make their customers their biggest fans, like Taylor has done.

Can you give a few examples of how that plays out weekly in the class?

One of my ideas for our marketing unit was a deep dive into Taylor’s style evolution over her self-proclaimed eras, and how that has reinforced her principles of relatability and world building. While style was a more subtle signal that built up over time, I’ve also enjoyed speaking about her direct power moves. Last night, for instance, we discussed how Taylor negotiated her contract with AMC Theatres and took hold of the reins for the Eras Tour film project. She financed the film and received 57% of the movie profits. To me, that was her learning from the mistake she made when she was younger, when she signed over the masters to her music.

In business school, students study the importance of connection in building an authentic brand. How has Taylor become a master at that?

Taylor’s songwriting stands out on two primary levels. The first is that she puts her insecurities and struggles out there, emotionally stripping herself through art. The second is that she vividly weaves those vulnerabilities into stories. Unique structures, sonic devices, and figurative language add layered complexities to these stories that ensure that they are highly talked about among consumers as a hot commodity. These elements of songwriting craft also tailor each product to match the message it is sending, which strengthens its value to consumers. She’s able to create a dynamic, so people continue to feel like they can relate to her. She really knows her audience, and her songs cover every part of her ideal listener’s life.

What does Taylor teach us about how to lead?

Taylor’s grandmother, Marjorie, said it best: “Never be so kind, you forget to be clever / Never be so clever, you forget to be kind.”

Taylor shows us how to balance a good heart with strategic design. We bring it up in class—the bonuses that she gives her team and the ways that she gives back to the community. Philanthropy happens to also be a tax write off for her, but that isn’t a bad thing. I think people know when a brand is doing something that feels inauthentic, and that isn’t the case with Taylor.

I think people know when a brand is doing something that feels inauthentic, and that isn’t the case with Taylor.

Taylor has so much power. How do you see her using it to uplift women’s voices, big and small?

Taylor has spoken extensively on how navigating the industry as a woman is different than as a man, which she writes about in “The Man” and “mad woman.”

She wears clothes from small, women-owned businesses, which have seen huge jumps in customers and traction.

But arguably one of the biggest ways that Taylor has amplified women’s voices is when she was a victim of sexual assault and ended up suing her assaulter for a symbolic one dollar. For many women, especially young fans, hearing a beloved figure speak so openly about that emotional damage not only acknowledges their pain, but also models speaking out against intolerable behavior that has become normalized in our society.

I have to ask about her dating Travis Kelce and what that has done for her brand.

The question should be what dating Taylor Swift has done for Travis Kelce’s brand. We’ll discuss her influence in the NFL in class and perhaps the perceptions that come with being in a high-profile relationship.

How much longer do you think that Taylor will continue reinventing herself as an artist? Do you think she will be like Madonna, touring in her 60s?

A lot of artists, once they feel like they’ve reached a certain point, go off the grid. I don’t quite know, but I know this: Taylor will always be a songwriter. She’s even said that she would consider writing songs for other people at some point. She cites songwriting as her lifeline, passion, and purpose—singing and performing are extensions of that.

Note: Bushala and her team will present at the annual Berkeley Haas Alumni Conference on April 27. Registration is open.  

Online images may be turning back the clock on gender bias, research finds

A paper published today in the journal Nature finds that online images show stronger gender biases than online texts. Researchers also found that bias is more psychologically potent in visual form than in writing.

An illustration of a hand holding a phone with a beam of light shining one a woman's face. A larger verison of her face is ampified in the background.
Image copyright Solène Delecourt

A picture is worth a thousand words, as the saying goes, and research has shown that the human brain does indeed better retain information from images than from text.

These days, we are taking in more visual content than ever as we peruse picture-packed news sites and social media platforms. And much of that visual content, according to new Berkeley Haas research, is reinforcing powerful gender stereotypes.

Through a series of experiments, observations, and the help of large language models, professors Douglas Guilbeault and Solène Delecourt found that female and male gender associations are more extreme among images retrieved on Google than within text from Google News. What’s more, while the text is slightly more focused on men than women, this bias is over four times stronger in images.

“Most of the previous research about bias on the internet has been focused on text, but we now have Google Images, TikTok, YouTube, Instagram—all kinds of content based on modalities besides text,” says Delecourt. “Our research suggests that the extent of bias online is much more widespread than previously shown.”

Not only is online gender bias more prevalent in images than in text, the study revealed, but such bias is more psychologically potent in visual form. Strikingly, in one experiment, study participants who looked at gender-biased images—as opposed to those reading gender-biased text—demonstrated significantly stronger biases even three days later.

As online worlds grow more and more visual, it’s important to understand the outsized potency of images, says Guilbeault, the lead author on the paper.

“We realized that this has implications for stereotypes—and no one had demonstrated that connection before,” Guilbeault says. “Images are a particularly sticky way for stereotypes to be communicated.”

The extent of bias–and its effects

To zero in on gender bias in online images, Guilbeault and Delecourt teamed up with co-authors Tasker Hull from Psiphon, Inc., a software company that develops censorship-navigation tools; doctoral researcher Bhargav Srinivasa Desikan of Switzerland’s École Polytechnique Fédérale de Lausanne (now at IPPR in London); Mark Chu from Columbia University; and Ethan Nadler from the University of Southern California. They designed a novel series of techniques to compare bias in images versus text, and to investigate its psychological impact in both mediums.

First, the researchers pulled 3,495 social categories—which included occupations like “doctor” and “carpenter” as well as social roles like “friend” and “neighbor”—from Wordnet, a large database of related words and concepts.

To calculate the gender balance within each category of images, the researchers retrieved the top hundred Google images corresponding to each category and recruited people to classify each human face by gender.

Measuring gender bias in online texts was a trickier proposition—though one perfectly suited for fast-evolving large-language models, which noted the frequency of each social category’s occurrence alongside references to gender in Google News text.

The researchers’ analysis revealed that gender associations were more extreme among the images than within the text. There were also far more images focused on men than women.

Sticky images

The experimental phase of the study sought to illuminate the impacts that biases in online images have on internet users. The researchers asked 450 participants to use Google to search for apt descriptions of occupations relating to science, technology, and the arts. One group used Google News to find and upload textual descriptions; another group used Google Images to find and upload pictures of occupations. (A control group was assigned the same task with neutral categories like “apple” and “guitar.”)

After selecting their text- or image-based descriptions, the participants rated which gender they most associated with each occupation. Then they completed a test that asked them to quickly sort various words into gender categories. The test was administered again after three days.

The participants who worked with the images displayed much stronger gender associations compared to those in the text and control conditions—even three days later.

“This isn’t only about the frequency of gender bias online,” says Guilbeault. “Part of the story here is that there’s something very sticky, very potent about images’ representation of people that text just doesn’t have.”

Interestingly, when the researchers conducted their own online survey of public opinion—and when they looked at data on occupational gender distributions reported by the U.S. Bureau of Labor Statistics—they found that gender disparities were much less pronounced than in those reflected in Google images.

Opening doors to new research

Delecourt and Guilbeault say they hope their findings lead to a more serious grappling with the challenges posed by embedded bias in online images. After all, it’s relatively easy to tweak text to be as neutral as possible, whereas images of people inherently convey racial, gender, and other demographic information.

Guilbeault notes that other research has shown that gender biases in online text have decreased, but those findings may not reveal the whole story. “In images we actually still see very prevalent widespread gender bias,” he says. “That may be because we haven’t really focused on images in terms of this movement towards gender equality. But it could also be because it’s just harder to do that in images.”

Guilbeault and Delecourt are already at work on another project in this vein to examine gender-age bias online using many of the same techniques. “Part of the reason this paper is so exciting is that it opens the door to many, many other types of research—into age or race, or into other modalities, like video,” Delecourt says.

Watch a video explaining the research:

Read the paper:

“Online Images Amplify Gender Bias
Nature, February 14, 2024

Authors:

  • Douglas Guilbeault: Haas School of Business, University of California, Berkeley (corresponding author)
  • Solène Delecourt: Haas School of Business, University of California, Berkeley
  • Tasker Hull: Psiphon Inc.
  • Bhargav Srinivasa Desikan: Ecole Polytechnique Federale de Lausanne
  • Mark Chu: Columbia University
  • Ethan Nadler: University of Southern California
This project was funded with grants from:

Contacts:

Douglas Guilbeault, corresponding author, [email protected]

Laura Counts, Berkeley Haas media relations: [email protected]