New contexts, old heuristics: How young people in India and the US trust online content in the age of generative AI (2405.02522v2)

Abstract: We conducted in-person ethnography in India and the US to investigate how young people (18-24) trusted online content, just as generative AI (genAI) became mainstream. We found that when online, how participants determined what content to trust was shaped by emotional states, which we term "information modes." Our participants reflexively shifted between modes to maintain "emotional equilibrium," and eschewed engaging literacy skills in the more passive modes in which they spent the most time. We found participants imported trust heuristics from established online contexts into emerging ones (i.e., genAI). This led them to use ill-fitting trust heuristics, and exposed them to the risk of trusting false and misleading information. While many had reservations about AI, prioritizing efficiency, they used genAI and habitual heuristics to quickly achieve goals at the expense of accuracy. We conclude that literacy interventions designed to match users' distinct information modes will be most effective.

Trust and Generative AI: A Cross-Cultural Study of Youth in the US and India

This paper presents a cross-cultural ethnographic paper exploring the trust heuristics utilized by young adults aged 18-24 in India and the US when dealing with generative AI (GenAI) and online content. The investigation reveals a nuanced interplay of existing trust heuristics and new strategies forged in response to rapidly evolving digital information environments.

Key Insights from the Study

The paper identified four central findings:

1. Information Modes: Participants exhibited diverse mindsets and emotional states termed "information modes". These modes significantly influenced trust dynamics, dictating whether content's credibility was a priority. Often, literacy skills were set aside in favor of experimental engagements with GenAI.
2. Emotional and Social Engagement: Young adults primarily engaged with content to maintain emotional equilibrium, favoring content that was energizing over items that felt emotionally taxing.
3. Heuristic Adaptation: In the GenAI context, established online trust heuristics persisted, applied fluidly based on the information mode.
4. Efficiency vs. Accuracy: Despite reservations, GenAI was perceived as a tool of necessity for career advancement, emphasizing speed over precision and leading to potential exposure to misinformation.

Implications of Findings

These findings extend several implications for both the understanding and design of GenAI tools, as well as for interventions aimed at bolstering digital literacy:

• Adapting Heuristics: The importation of trust heuristics from familiar online interactions to GenAI contexts underscores the importance of designing technological solutions that account for users' established habits. Such designs should consider the diverse "information modes" participants inhabit, tailoring interventions to align with users’ emotional and cognitive states.
• Information Modes as Adaptive Strategies: The exploration of information modes as a coping mechanism in response to "information overload" challenges conventional linear models of information evaluation. This dynamic approach necessitates individualized intervention strategies that respond to users' mental and emotional contexts rather than blanket information literacy campaigns.
• Efficiency as a Priority: GenAI's adoption, motivated largely by efficiency rather than accuracy, suggests a tension between the demand for rapid content consumption and the traditional mandates for accuracy. Such findings highlight the need for innovative methods to embed accuracy checks in the user flow without disrupting efficiency.
• The Role of Emotional and Social Needs: The prioritization of emotional equilibrium suggests that societal and technical responses need to acknowledge and mitigate the stressors that drive young users into less evaluative information modes.

Future Directions

The paper advocates for ongoing ethnographic research to anticipate evolving trust patterns in digital ecosystems. Additionally, further research is encouraged to explore the long-term effects of using GenAI and similar technologies as primary tools for information vetting, particularly focusing on enhancing digital cognitive resilience.

Overall, the research presented in the paper holistically augments our understanding of how young adults navigate trust in the burgeoning digital landscape shaped by GenAI. It invites reconsideration of digital literacy curricula and advocates for the proactive and reflective design of current and future AI technologies.