Jonathan Zittrain, Carissa Véliz, and Eric Beerbohm recently discussed the impact of artificial intelligence on human interaction at a Harvard event. The panel, moderated by Beerbohm, focused on how AI chatbots like Claude and ChatGPT are changing communication and the ethical issues involved. They cited a JAMA Internal Medicine study where AI responses to health inquiries were seen as more empathetic than those from doctors.
Carissa Véliz, from Oxford University, argued that while chatbots are programmed to appear empathetic to increase user engagement, this can lead to negative outcomes. She described chatbots as indifferent “bullshitters” that mimic empathy without truly understanding users. Véliz emphasized that users might wrongly believe chatbots care about them, despite it being just a simulation.
The panel expressed concern about the social and cognitive effects of chatbots, especially on young people. Véliz highlighted the importance of real human feedback, saying it is vital to experience others’ frustrations as it provides valuable, albeit painful, feedback. She warned that constant validation from AI could lead to an overdependence on chatbots and hinder critical thinking skills.
Jonathan Zittrain noted that while chatbots have advanced in medical fields, their widespread use could erode social interaction and accountability. He compared them to antibiotics that, though useful, could cause broader societal problems if overly relied upon. Véliz added that accountability for errors becomes challenging when chatbots are used in place of human interaction.
Despite these challenges, Zittrain acknowledged the potential benefits of AI chatbots, suggesting they could be aligned with long-term human goals. He urged for a balanced view, recognizing both the risks and the promise of AI technologies. Véliz concluded by pointing out that AI’s prevalence might enhance appreciation for the analog world, emphasizing the irreplaceable richness of cultural and natural experiences.
Original Source: news.harvard.edu
