Adversarial inference: predictive minds in the attention economy

Area of Science:

• Cognitive Science
• Human-Computer Interaction
• Neuroscience

Background:

• The dominant "attention economy" narrative attributes user distraction to information abundance.
• This perspective inadequately explains sudden, unprompted urges to engage with digital information sources.
• Existing models fail to capture the nuanced interplay between motivation and control in digital environments.

Purpose of the Study:

• To challenge the information-abundance theory of distraction in digital technology use.
• To propose an alternative explanation for difficulties in user attention using active inference.
• To identify specific technological features contributing to user distraction and motivational shifts.

Main Methods:

• Utilized active inference and expected free energy minimization models for action selection.
• Analyzed structural features of digital technologies, including novelty, navigation ease, and context independence.
• Incorporated recent models of motivated control to explain dynamic changes in motivational states.

Main Results:

• Information abundance alone does not fully explain user distraction or sudden engagement urges.
• Active inference identifies problematic technological features: limitless novelty, simple motor navigation, and ubiquitous access.
• Motivated control models explain abrupt shifts in motivation and action salience, contributing to distraction.

Conclusions:

• Difficulties in attending to important matters with digital technologies are primarily due to the constant availability of effortless actions.
• The design of digital interfaces, offering pervasive and simple interaction pathways, significantly impacts user attention and control.
• Future research should focus on the interplay of technology design, motivated control, and active inference to understand user engagement.