Why misinformation persists

While misinformation may seem particularly pervasive in the digital age, it is far from a new phenomenon. Technology has amplified its reach and urgency, but the tendency to create and spread false information long predates the internet.

Yotam Ophir, associate professor of communication and director of UB's Media Effects, Misinformation, and Extremism (MEME) Lab, has spent years studying misinformation.

"Misinformation isn't a new problem," Ophir says. "But AI is changing the speed, scale, and accessibility of content creation in ways that are transforming the information environment."

As artificial intelligence becomes increasingly integrated into everyday life, Ophir is helping uncover how false and misleading information spreads, why people believe it, and how emerging technologies may be changing the information environment. One major focus of the MEME Lab is understanding how scientific issues become politicized. Topics such as vaccinations, climate change, and public health were once viewed largely through scientific lenses but have increasingly become markers of political identity. According to Ophir, computational methods can help pinpoint when that shift occurs and how it spreads.

"Once a scientific topic becomes politicized, the weight of evidence often becomes secondary to what that issue represents about a person's social or political group," he says.

Ophir's work focuses on computational communication science, which is an interdisciplinary field that combines machine learning, natural language processing, and network analysis to study communication at a scale that would have been impossible just a decade ago. Using machine learning tools, Ophir and his lab identify recurring themes, track shifts in public discourse, and map how narratives move across digital communities.

"When we look at misinformation today, we're analyzing massive volumes of digital content that have accumulated over years or even decades," Ophir says.

These methods allow the researchers to investigate not only individual falsehoods, but also larger patterns in how information ecosystems operate.

While Ophir has long used machine learning to analyze data, public interest in AI surged following the release of ChatGPT and other generative AI systems. In response, students began asking new questions about how people understand, trust, and adopt AI technologies. Those questions have since become a growing area of investigation within the MEME Lab.

"From the moment I started learning about mass media and its effects, I was immediately drawn to questions of deception, propaganda, and misinformation," Ophir says. “At first I studied it primarily in the context of health, but while working on my 2025 book, Misinformation and Society, my interest slowly shifted more towards political misinformation and conspiracy theories, in general, and among extremist groups in particular."

His early work examined misleading claims surrounding tobacco products, including e-cigarettes and so-called organic tobacco. During the Zika virus outbreak, he began investigating misinformation during public health crises, a research area that became increasingly important during the COVID-19 pandemic. As debates over vaccines and public health measures became deeply political, Ophir's research expanded into the broader relationship between misinformation, political identity, and extremist movements.

Despite the sophistication of machine learning tools, Ophir’s research ultimately returns to a fundamental fact about people.

“Humans have not evolved to be truth-seekers,” he says, and explains that evolution favored survival and group cohesion, not accuracy. For most of human history, following social norms and responding quickly to threats mattered more than verifying information sources.

Current projects examine public perceptions of AI, the role of recommendation algorithms in shaping what people see online, and the ways AI-generated content may influence public discourse. Ophir emphasizes that AI should not be viewed in isolation but as part of a longer history of communication technologies that have transformed society.

"Many people felt unprepared for how quickly generative AI entered everyday life," he says. "But history shows that we've experienced similar moments of disruption with previous technologies. Understanding that history helps us respond to the new threats AI poses to information integrity and the resilience of democracy thoughtfully, rather than react out of fear."

Ophir’s research also shapes his teaching. In the classroom, he encourages students to question assumptions, including his own. For students interested in AI, data science, or communication, he stresses that technical expertise alone is not enough.

“You can’t understand AI without understanding the people, institutions, and incentives behind it,” he says.

These days, Ophir is coauthoring a book on AI with Joaquin Carbonara, PhD, a computer scientist at Buffalo State, and Danielle Ophir, a human factors PhD student at UB, to try and provide a comprehensive answer to some of these questions.