作者: Max M. Louwerse , Arthur C. Graesser , Danielle S. McNamara , Shulan Lu
DOI: 10.1002/ACP.1527
关键词:
摘要: SUMMARY Conversational agents are becoming more widespread in computer technologies but there has been little research how humans interact with them. Two eye tracking studies investigated distribute gaze towards conversational complex tutoring systems. In Study 1, participants interacted the single-agent system AutoTutor. Fixation times showed that agent received most attention throughout interaction, even when display size was statistically controlled. 2, iSTART. Fixations were on relevant these spoke. Both provided evidence regard animated as partners communication process. Copyright# 2008 John Wiley & Sons, Ltd. Embodied (ECAs) characters emulate human behaviour and communication. There a boom of environments vary from conventional web systems to advanced virtual worlds. A fundamental question arises interpret agents. Are ECAs distracting artefacts or human-like partners? Answers questions require an in-depth analysis humans’ perceptions, interactions The present designed address questions. Some have suggested users they would (Reeves Nass, 1996), presumably because computer-generated demonstrate many same multi-modal properties do face-to-face conversation (Cassell, Sullivan, Prevost, Churchill, 2000). Studies cognitive psychology, social artificial intelligence education effects user impressions, comprehension learning gains (Atkinson, 2002; Baylor Ryu, 2003; Graesser, Moreno, Marineau, Adcock, Olney, Person, 2003). shown conditions which interacting displayed monitor benefit interfaces than