Computing Reviews
Today's Issue Hot Topics Search Browse Recommended My Account Log In
Review Help
Search
Tariff agent: interacting with a future smart energy system at home
Alan A., Costanza E., Ramchurn S., Fischer J., Rodden T., Jennings N. ACM Transactions on Computer-Human Interaction23 (4):1-28,2016.Type:Article
Date Reviewed: Nov 2 2016

Are we ready to fully cede control to intelligent agents, such as self-driving cars? Or would we prefer to have the option of taking over control when we deem it appropriate?

In their paper, the authors investigate the use of intelligent agents for the selection of suitable tariffs (subscription plans) for home electricity. The assumption is that consumers can switch between different plans on a daily basis, based on their preferences (renewable versus fossil fuel), usage patterns, and price fluctuations. While some of these factors are fairly stable, others change frequently, and users may benefit financially by switching plans. Since these decisions may involve simple but somewhat tedious deliberations on a regular basis, intelligent agents can offer various degrees of assistance to consumers.

Participants in the experiments interacted with an agent-based system on three levels of autonomy: the default level, suggestion-only, where the agent provides information about the most suitable tariff for the next day; semi-autonomous, where the agent switches to the best predicted tariff and informs the user via text message; and fully autonomous, where the agent does the switch without informing the user. The experiments were conducted in the homes of participants, utilizing devices provided by the investigators. Participants also received moderate financial compensation reflecting their plan choices to provide realistic incentives.

While all participants appreciated the notifications, about half decided to switch to the semi-autonomous mode, mostly for convenience reasons. None of them switched to the fully autonomous mode, possibly due to the limited time period of the experiment. Based on data analysis and verbal feedback, users appreciated having a choice in the level of autonomy. Fully autonomous systems left them feeling out of control, while flexible autonomy helped to establish a level of trust in the system by providing a balance between monitoring system performance and delegating control.

While the context of these experiments is clearly quite different from self-driving vehicles, the option to maintain some control over autonomous systems was a clear preference among the participants. For me, gradually ceding control to an agent in a similar scenario certainly would be an option. I’d certainly be much more reluctant to use a vehicle without a steering wheel or brakes, where taking back control clearly is not an option.

Reviewer:  Franz Kurfess Review #: CR144893 (1702-0159)
Bookmark and Share
  Featured Reviewer  
 
User Interfaces (H.5.2 )
 
Would you recommend this review?
yes
no
Other reviews under "User Interfaces": Date
Computers as theatre
Laurel B., Addison-Wesley Longman Publishing Co., Inc., Boston, MA, 1991. Type: Book (9780201510485)
Dec 1 1991
Intelligent user interfaces
Sullivan J., Tyler S., ACM Press, New York, NY, 1991. Type: Book (9780201503050)
Feb 1 1993
Electronic performance support systems
Gery G., Weingarten Pub., Inc., Boston, MA, 1991. Type: Book (9780961796815)
Sep 1 1992
more...

E-Mail This Printer-Friendly
Send Your Comments
Contact Us
Reproduction in whole or in part without permission is prohibited.   Copyright 1999-2024 ThinkLoud®
Terms of Use
| Privacy Policy