Attention please! Next-generation e-learning is here

Recent trials of two new software platforms based on this new approach show substantial promise in overcoming one of the biggest problems that has dogged e-learning: how to keep students motivated and attentive. The platforms, developed in the AtGentive project, are designed to aid students in the classroom and to help them continue learning and collaborating long after classroom sessions have ended.

“The first generation of e-learning platforms focused on replicating online the classroom model of teaching, but this approach has not been all that successful,” explains Thierry Nabeth, the coordinator of AtGentive at INSEAD’s Centre for Advanced Learning Technologies in France. “The biggest problem is that students often lack motivation both inside and outside of the classroom, and fail to dedicate their attention to the learning programme.”

In an effort to overcome that problem, the AtGentive researchers incorporated artificial agents and social networking into their approach toward e-learning, employing, in the case of one of the platforms, similar techniques to those that have made websites, such as Facebook, so popular as a means of staying in touch with friends, relatives and colleagues.

Keeping a (virtual) eye on the class…
“Artificial agents are autonomous entities that observe users’ activities and assess their state of attention in order to intervene so as to make the user experience more effective,” Nabeth says. “The interventions can take many forms, from providing new information to the student, guiding them in their work or alerting them when other users connect to the platform.”

In an e-learning context, the agents provide a smart form of proactive coaching for students, assessing, guiding and stimulating them. For example, an artificial agent can alert a student when an article they have posted elicits the attention of other users, or when they receive feedback on their input into a collaborative learning task. In a classroom environment, an artificial agent embodied as an animated character spots students who are not interacting with the system and probably not paying attention. The avatar helps by trying to “wake them up.”

“For collaborative learning to be effective, it is important for people to know how their input is being received by others and whether what they are working on is of interest to other people,” Nabeth says, noting that in essence the team have adapted renowned psychologist Albert Bandura’s concept of self-efficacy to a collaborative context.

It is also important to minimise distractions. For that reason, the agents designed in the EU-funded AtGentive monitor what students are doing and only intervene when doing so will not unnecessarily take their attention away from the task at hand.

The results of trials involving the two AtGentive platforms show how effective the approach can be. One, AtGentSchool, which was tested in a primary school setting in the Czech Republic, helps maintain students’ attention in the classroom using avatars to interact with students on their PCs. The other, AtGentNet, is a social network platform incorporating artificial agents that has been used in a training programme for international trade managers run by the Swedish Trade Council (STC), a project partner.

“AtGentSchool has a more short-term (real-time) and individual focus on managing attention, while AtGentNet is more long-term (asynchronous), and focused on the group and on social attention,” Nabeth says.

More interaction, more motivation, more attention
In the AtGentSchool pilot, pupils stimulated by “attention aware” artificial agents have shown a higher level of satisfaction and motivation. However, whether the learning process is more effective is a question that will require future investigation, says Nabeth.

According to the project coordinator, trade managers who used a full-featured version of AtGentNet for collaborative learning tasks, in the weeks and months in between seminars, interacted more frequently and showed more motivation than two control groups who used a version with reduced features and who used a legacy e-learning platform.

“The results would need to have a bigger sample to be more conclusive but they suggest we are on the right track toward more effective e-learning,” Nabeth tells ICT Results.

AtGentNet continues to be used by STC and the project partners are rolling it out as a service for both e-learning applications and for research. They are working closely with FIDIS, a network of excellence addressing related issues in the information society.

“[Because our platform] monitors everything users do and shares information about them with other users, [it] will be used to research issues such as privacy and identity in social networks,” Nabeth notes.

Dutch firm Ontdeknet, whose artificial characters are the visible face of AtGentSchool, is continuing to develop the platform and is incorporating it into its product line as an e-learning system for children.

In Nabeth’s view, the AtGentive approach will help end the “shallow” experience of many e-learning systems developed to date, while deepening our understanding of what it takes to hold people’s attention and keep them motivated.