Conversational Agents (Feb. 21, 2009)
In the article, “When sex, drugs, and violence enter the classroom: conversations between adolescents and a female pedagogical agent,” the authors express surprise regarding the students’ abuse of the conversational agent. However, personally, I am not surprised in the slightest! Drawing from my experiences with pre-teens and teens, I’d argue that adolescent misuse of technological tools is rampant. For instance, recently, my friend told me that her thirteen year-old niece had been receiving sexually abusive text messages from a number of boys in her classroom. Afraid or embarrassed to tell her parents, the niece asked her aunt for advice. When my friend told me this story, we both remarked on the tendency for digital communication devices to lessen human inhibition (regardless of the person’s age). Why do typed, satellite-fueled messages seem to encourage regular dismissal of social norms?
The Sex, Drugs, and Violence article predicts that the abuse/misuse of conversational agents likely relates to the anonymity cyber space affords. Yet, this explanation can’t be applied to digital conversations like cell-phone texting in which the speakers are known. I’d venture to guess that the misuse of conservational agents connects to excessive teenage viewing/use of violent and sexually explicit video games, television shows, and movies. As learning theorists link the form of learning to the content of what is learned, perhaps teenagers link digital communication to the explicit content they view via digital means. In this way, cyberspace and text messaging become spaces where conversations about sex, drugs, and violence are acceptable and inconsequential.
The articles’ application of Media Equation Lens and Asimov’s Laws of Robotics provides insight into student relationship and interaction with conversational agents specifically. Velestianos, et. al. writes, “The media equation argues that human treat media as if they are also human, in essence interacting with media in the same way that humans would interact with each other” (293). As the Doering article discusses, people enjoy the human element of the conversational agent. Yet, as Asimov’s Laws of Robotics describes, “These laws position the human as an authoritative figure and the robot in a subordinate role. If users perceive robots (and by extension, virtual characters) to be their subordinates and servants, human-agent interactions may be influenced by such a power differential” (294). Interestingly, we see this paradox- dual perception of conversational agent as human and as robot- played out in the detailed students’ responses. In both articles, students converse with the agent on non-task (social) topics when the agent failed to give student the appropriate, task-related response. Students realize that the agent is not a human and, therefore, students feel comfortable asserting dominance and abuse.
In my opinion, it seems that students are interested in the conservational agents, but the agents fail to increase student learning. In nearly all the situations outlined, students spent more time conversing about social topics and abusing the agents than actually using the conservational agents for the intended purposes (assignment related inquiry, etc). However, I think the potential for increased learning is best described in the Doering article, “We have seen a move toward learning with technology rather than from technology” (268). Since students enjoy using the conversational agents, but the agents are incapable of higher level thinking, I’d consider incorporating non-expert agents in my teaching. If the agent and the student worked as collaborators rather than expert/non-expert, the abuse might decrease. (Although, in all honesty, I find that teenagers can be really quite perverse regardless...especially in a no-stakes situation, the language will get explicit.)
Ideas for classroom activities with conversational agents:
-Non-expert situations…
-Students create pedagogical agent of themselves…describing a task.
-Students create pedagogical agent of a literary character.
The Sex, Drugs, and Violence article predicts that the abuse/misuse of conversational agents likely relates to the anonymity cyber space affords. Yet, this explanation can’t be applied to digital conversations like cell-phone texting in which the speakers are known. I’d venture to guess that the misuse of conservational agents connects to excessive teenage viewing/use of violent and sexually explicit video games, television shows, and movies. As learning theorists link the form of learning to the content of what is learned, perhaps teenagers link digital communication to the explicit content they view via digital means. In this way, cyberspace and text messaging become spaces where conversations about sex, drugs, and violence are acceptable and inconsequential.
The articles’ application of Media Equation Lens and Asimov’s Laws of Robotics provides insight into student relationship and interaction with conversational agents specifically. Velestianos, et. al. writes, “The media equation argues that human treat media as if they are also human, in essence interacting with media in the same way that humans would interact with each other” (293). As the Doering article discusses, people enjoy the human element of the conversational agent. Yet, as Asimov’s Laws of Robotics describes, “These laws position the human as an authoritative figure and the robot in a subordinate role. If users perceive robots (and by extension, virtual characters) to be their subordinates and servants, human-agent interactions may be influenced by such a power differential” (294). Interestingly, we see this paradox- dual perception of conversational agent as human and as robot- played out in the detailed students’ responses. In both articles, students converse with the agent on non-task (social) topics when the agent failed to give student the appropriate, task-related response. Students realize that the agent is not a human and, therefore, students feel comfortable asserting dominance and abuse.
In my opinion, it seems that students are interested in the conservational agents, but the agents fail to increase student learning. In nearly all the situations outlined, students spent more time conversing about social topics and abusing the agents than actually using the conservational agents for the intended purposes (assignment related inquiry, etc). However, I think the potential for increased learning is best described in the Doering article, “We have seen a move toward learning with technology rather than from technology” (268). Since students enjoy using the conversational agents, but the agents are incapable of higher level thinking, I’d consider incorporating non-expert agents in my teaching. If the agent and the student worked as collaborators rather than expert/non-expert, the abuse might decrease. (Although, in all honesty, I find that teenagers can be really quite perverse regardless...especially in a no-stakes situation, the language will get explicit.)
Ideas for classroom activities with conversational agents:
-Non-expert situations…
-Students create pedagogical agent of themselves…describing a task.
-Students create pedagogical agent of a literary character.
I was also not surprised that the students reacted sexually to the agents. In a lot of schools they are talking to each other that way, and thats not even anonymous. With the high amount of sexual content in the media, students are more and more outwardly curious about sex, and it showed in that study. Disturbing though, huh?
ReplyDeleteWe get instances of the harassing texts/emails/chats here at the junior high quite often. You know, I found the perverse chat with the conversational agents disturbing, but not shocking because I've seen students writing things like that in textbooks!
ReplyDelete