Search for: All records

Effective human-AI collaboration hinges not only on the AI agent’s ability to follow explicit instructions but also on its capacity to navigate ambiguity, incompleteness, invalidity, and irrelevance in communication. Gricean conversational and inference norms facilitate collaboration by aligning unclear instructions with cooperative principles. We propose a normative framework that integrates Gricean norms and cognitive frameworks—common ground, relevance theory, and theory of mind—into large language model (LLM) based agents. The normative framework adopts the Gricean maxims of quantity, quality, relation, and manner, along with inference, as Gricean norms to interpret unclear instructions, which are: ambiguous, incomplete, invalid, or irrelevant. Within this framework, we introduce Lamoids, GPT-4 powered agents designed to collaborate with humans. To assess the influence of Gricean norms in human- AI collaboration, we evaluate two versions of a Lamoid: one with norms and one without. In our experiments, a Lamoid collaborates with a human to achieve shared goals in a grid world (Doors, Keys, and Gems) by interpreting both clear and unclear natural language instructions. Our results reveal that the Lamoid with Gricean norms achieves higher task accuracy and generates clearer, more accurate, and contextually relevant responses than the Lamoid without norms. This improvement stems from the normative framework, which enhances the agent’s pragmatic reasoning, fostering effective human-AI collaboration and enabling context-aware communication in LLM-based agents. 

Synchronous, face-to-face interactions such as brainstorming are considered essential for creative tasks (the old normal). However, face-to-face interactions are difficult to arrange because of the diverse locations and conflicting availability of people—a challenge made more prominent by work-from-home practices during the COVID-19 pandemic (the new normal). In addition, face-to-face interactions are susceptible to cognitive interference. We employ crowdsourcing as an avenue to investigate creativity in asynchronous, online interactions. We choose product ideation,a natural task for the crowd since it requires human insight and creativity into what product features would be novel and useful. We compare the performance of solo crowd workers with asynchronous teams of crowd workers formed without prior coordination. Our findings suggest that, first, crowd teamwork yields fewer but more creative ideas than solo crowdwork. The enhanced team creativity results when (1) team workers reflect on each other’s ideas, and (2) teams are composed of workers of reflective, as opposed to active or mixed, personality types. Second, cognitive interference, known to inhibit creativity in face-to-face teams, may not be significant in crowd teams. Third, teamwork promotes better achievement emotions for crowd workers. These findings provide a basis for trading off creativity, quantity, and worker happiness in setting up crowdsourcing workflows for product ideation.