Cognitive Bias and Demand Response
The transition to a smarter electric grid introduces several challenges that could threaten the grid’s reliability, such as new kinds of electrical loads (i.e. electric vehicles) and less-predictable energy sources (wind, solar, etc.). Traditionally, utilities have turned to demand response (D-R) programs, which use financial incentives to motivate consumers to redistribute their electric load, to manage the grid during peak periods. On paper, using demand response is a win for both utilities and customers: The former group mitigates peak electric loads and makes the grid more reliable, while consumers see a reduction in their energy costs.
The problem here is that demand response programs make several critical assumptions about consumers:
- Consumers will draw accurate conclusions about the state of the world based on a) the data they receive, and b) the method by which the utility provides it.
- Consumers will make rational decisions based on this data—in this case, a decision to minimize their energy costs.
- Consumers will then use the given technology to correctly implement their decisions.
These assumptions, however, are challenged by a large body of psychology work, which offers a list of cognitive biases that can manipulate the mind’s decision-making process in predictable, repeatable ways. If we do not take into account how consumers actually behave (and how cognitive biases influence their behavior), not only will any demand response program or technology we deploy not achieve its goal, but an attacker could put the grid in a precarious position by using these biases to manipulate customer behavior in a dangerous manner (for example, cause consumers to schedule much of their electric loads during peak periods).
In this task, a computer scientist (who has identified other areas where cognitive bias has disrupted computer security) has partnered with a psychologist in guiding a series of students to produce an Amazon Mechanical Turk experiment to determine what cognitive bias issues might be at play in smart grid demand response. The experiment will go live when the students return from summer break.