With its exponential growth, many people worry that artificial intelligence will render humans obsolete. But according to a variety of research by faculty at the University of South Florida, there are many instances where AI can’t compete with the human touch.
Humans and AI tools have different skill sets and shouldn’t be pitted against one another. While AI excels at synthesizing large amounts of data, humans shine in the supervisory role.
Dipayan Biswas, a professor of marketing at USF compared AI to fire — an amazing tool, but dangerous when out of control.
“When AI's power is harnessed as an assistant, as a servant, it's amazing. Whenever we let the guardrails go, when ever AI takes on the master's role, that's when things get really messy,” he said.
‘Humans are needed’
They build [AI apps] with an agenda where warmth becomes a slot machine — affectionate enough to hold you, needy enough that you can't leave, so that you keep paying for the rest of your life.
Triparna de Vreede, director of the Behavioral AI Research Lab at the University of South Florida’s Muma College of Business
Humans shouldn’t worry about being replaced by AI any time soon.
“Unless we can create AI that truly can love, truly be present, and truly can create connection, humans are needed. There’s no replacement for that,” said Triparna de Vreede, director of the USF Behavioral AI Research Lab.
Her background as a psychologist heavily influences her AI research as assistant professor of information systems at USF.
De Vreede co-authored the 2025 study, “Human + AI Therapy: A Hybrid Mental Health Approach,” that explored pairing AI tools with human therapists. She said the point of AI in the care field is to allow people to thrive even when there is a shortage of human caretakers and basic human connection in daily life.
“I want to see if we can use AI to not replace human witnessing, not replace human connections, but to amplify it,” she said.
For example, the study examines an AI assistant app that mental health patients use to share how they are doing throughout the week. Then, the therapist can use that information to better prepare for the patient’s weekly appointment.
“Any AI that simply replaces humans is going to exacerbate the problem,” said de Vreede, adding that most corporations have a primary goal of keeping people subscribed, which translates to profit.
“They build [AI apps] with an agenda where warmth becomes a slot machine — affectionate enough to hold you, needy enough that you can't leave, so that you keep paying for the rest of your life,” she added.
Rather, de Vreede said that AI apps should be built to help people create stronger connections with other humans. Without that human feedback, users of AI mental health apps are becoming less connected to the real world.
“AI developers sitting in corporations have made it so that it will never disagree with you, always find logic in what you are saying, always be on your side — at some point, you will lean on it, no matter how smart you are,” she said. “And that is why AI ethics are so important.”
Human-designed products take the upper hand
When it comes to sustainability, having the human touch lends it more credibility.
Although AI can be more efficient than humans in designing sustainable products, people perceive products as more sustainable when told they were designed by humans, according to new research published in the Journal of the Academy of Marketing Science.
“Anything coming from AI is viewed with suspicion, and rightly so, because AI doesn't
have the concept of living or dying,” said Biswas, who co-authored the 2026 study,
“AI and Human Designers: How Consumers See the Genuine Care in Product Design as More
Sustainable.”
For technology products such as industrial machinery, AI is perceived more positively. “We just look strictly at efficiency,” he added, noting BMW has leaned into talking about its use of AI in its marketing. “They have embraced it, and it's working out well.”
Not so for Coca-Cola and Tommy Hilfiger, both of whom suffered a public backlash over their use of AI. “When it comes to food, beverages, clothing, and fashion design, the human touch is more valued,” Biswas said. “AI cannot smell. It has no concept of it's a nice scent, or if it's esthetically appealing. So [in] the sensory domain, humans tend to have the upper hand.”
Keeping a human in the loop
Preserving that human connection seems crucial in the hospitality sector, yet the rapid integration of service robots in hotel and restaurant settings is transforming employee roles, service delivery, and guest experiences.
In fact, hotel managers are more eager to adopt AI to counteract the industry’s staffing shortage, according to research published in the Journal of Hospitality and Tourism Technology.
While older customers tend to prefer human assistance, younger customers are happy that AI speeds up basic automated functions, said Luana Nanu, assistant professor of hospitality at USF and co-author of the study, “Investigating Customer Service and Engagement Levels in the Lodging Industry: High Touch to High Tech Conversational AI.”
“But when things get a little rougher, like cancellations, things that kind of strike a nerve. They really did not like talking to the AI about that,” Nanu said, noting that denying customers the option to speak to a human leads to frustration, dissatisfaction and lost business. “We call it ‘human in the loop’ – basically, have a person on standby for more complex requests.”
AI as a travel planner
Travelers have turned to AI to map out their next dream destination itinerary.
Consumers using AI as a de facto travel planner jumped six-fold in the past three years, from hovering at a modest 10 to 12% to a staggering 60 to 65%, according to Seden Doğan, an assistant professor of instruction at USF.
She co-authored the study, “Antecedents and Consequences of Travelers' Trust Towards Personalized Travel Recommendations Offered by ChatGPT”, published in 2023 in the International Journal of Hospitality Management.
Doğan cited four important reasons why people use AI for travel planning: personalization, speed, convenience and 24/7 access. “The more details you give, the more detailed itinerary you get from the generative AI, everything in just one minute,” she said.
Despite AI’s increase in popularity for travel planning across all generations, Doğan said human travel agents aren’t going extinct because travelers cannot blindly trust AI-generated itineraries.
“The reality is that generative AI hallucinates and gives you inaccurate information,” she said. “And what if something happens in the destination you are traveling? You cannot ask generative AI, ‘Hey, this happened. Help me to solve it.’ No, then you really need to have a conversation with a real person.”
And using AI as a replacement for expert human advice can be a recipe for disaster.
“The point of contention is how to interpret that information,” Biswas said. Although he uses AI as an assistant daily, he rejects about half of what it produces. “A lot of it is actually sloppy work. You need to have that position of knowledge to know whether the output is good versus not.”
When AI's power is harnessed as an assistant, as a servant, it's amazing. Whenever we let the guardrails go, whenever AI takes on the master's role, that's when things get really messy.
Dipayan Biswas, professor of marketing at the University of South Florida’s Muma College of Business
Undermining human growth
De Vreede’s research shows that people pride themselves on who they are, defined to a large extent by what they know.
“Now when AI comes into organizations… you can do many things faster, but at the end of the day, when you are looking at yourself, your sense of self will start eroding,” she said, noting it leads to disengagement.
She co-authored the 2026 study, “The Developmental Bypass: How AI Assistance Undermines Human Growth.”
Experts lose their sense of self more quickly than novices do, because the novices never had that expertise to lose in the first place. Hence, novices feel more productive when using AI in their tasks, she said.
Organizations should introduce AI differently depending on the employees’ experience level. “With experts, the AI shouldn't provide the answers. AI should ask the questions and help the person think it through faster, better,” de Vreede said. “Whereas with novices, AI should answer questions and provide a structured way of thinking so that they can think and become experts.”
She said people should learn to use AI for collaborative conversation. “Otherwise, the first place we go to is asking for solutions, not asking to be asked questions.”
Unfortunately, de Vreede said, AI is being used as a one-size-fits-all solution by many organizations. “People are unique.” she said. “And what is bringing out brilliance in one group of people is dominating another group.”
AI filling in the gaps
So how can AI be used in the most optimal way?
De Vreede said when an organization’s leaders truly
understand AI’s capabilities, they can have an honest
conversation with their human employees to understand
their responsibilities and what may be lacking.
Then, AI can create that needed support.
“All of a sudden, you have given them a blessing,” she
said. “They can now be the best people they can be
because instead of taking away their strengths, we are
filling their gaps.”
Her most recent paper, forthcoming in the Journal of
Management Information Systems, showed that human
performance improves when a human teammate excels
but worsens when an AI teammate excels.
She often gets asked what are humans to do in the age
of AI when AI is doing everything. Her answer is simple.
“We are supposed to become more human. That's what AI
will help us do, if we do it correctly.” ROI