Beyond Automation: How Researchers Are Unpacking the Human Side of Artificial Intelligence
Table of Contents
Artificial intelligence is rapidly transforming our world, often framed by discussions of automation, efficiency, and disruption. But for a cohort of researchers at the University of Massachusetts Amherst, AI represents a powerful lens for understanding fundamental aspects of the human experience – sound, trust, and even humor. Their work delves into how existing technologies reflect and reshape human values, and how people respond to machines that increasingly perceive and interact with the world around them.
The Interdisciplinary Approach to Understanding AI’s Impact
machine listening, exploring artificial intelligence through the study of sound. While much attention focuses on AI’s ability to generate audio, Paskar is interested in what happens before generation – the crucial process of sound recognition. “I’m interested in sound recognition that is fueled by artificial intelligence, machine learning,” she explains. “There is this term that people sometimes use, machine listening.”
Paskar’s work moves beyond the familiar categories of music and speech, focusing instead on “the rest of the sonic world” – environmental sounds, background noise, and nonhuman audio. She notes that humans currently excel at categorizing these nuanced sounds – distinguishing, such as, between a burglar and a cat – but that AI technology is rapidly advancing. Paskar sees her research as a way to raise awareness of the inherent assumptions embedded within technology. “Sound is realy under-looked… underrated,” she says. She also draws parallels between current anxieties surrounding AI and past reactions to new technologies,like the phonograph,suggesting that fears often reflect broader social shifts.
The proliferation of deepfakes – AI-generated videos and images – is raising critical questions about trust and perception. Doctoral candidate Shahnaz Bashir is analyzing audience reactions to these increasingly sophisticated forgeries, focusing on how viewers interpret authenticity, intent, and credibility across different cultural contexts.
Bashir’s research reveals that deepfakes are no longer limited to political disinformation. They are now appearing in advertising, social media influencing, and even as a means for individuals to cope with loss by creating AI-generated versions of deceased loved ones. Though, this widespread use is contributing to a growing sense of visual uncertainty. “It’s really difficult to tell a video apart from its fake version,” Bashir observes.”The technology is making it more refined and sophisticated.”
He cautions against relying solely on detection tools, warning that “there’s no forensics that can tell you what is and is not a deepfake,” and emphasizes the importance of verification and critical thinking. Bashir also highlights the danger of the “liar’s dividend,” where genuine footage can be dismissed as fake simply becuase deepfakes exist, underscoring the urgent need for enhanced media literacy.
Can AI Understand Humor Across Cultures?
Doctoral student Ibukun Filani approaches AI through the complex lens of humor, language, and culture. His research tackles a deceptively simple question: Can AI tell African jokes? Originally from Nigeria, Filani studies how AI systems generate humor based in African cultures, and why those attempts frequently enough fall flat.
Humor, he explains, is deeply rooted in shared knowledge, language, and social context.While computer-generated jokes have existed since the mid-1990s, african humor presents unique challenges for AI systems trained primarily on Western data.When jokes fail to land, Filani sees a reflection of the biases inherent in machine learning. “AI tends to just simply reproduce the colonial archive,which in itself is incomplete,” he observes. He argues that large language models learn not just language, but also genres – the frameworks through which we understand the world – and that this process often prioritizes Western perspectives. By using humor as a focal point, Filani’s work opens broader conversations about representation, inclusion, and cultural knowledge in the development of artificial intelligence, extending far beyond the realm of comedy.
