Brainrot: Deskilling and Addiction are Overlooked AI Risks

Cognitive decline from over-reliance is an overlooked AI safety concern: The paper identifies "deskilling" and atrophy of critical thinking skills as significant risks from depending too heavily on GenAI systems, yet these concerns are rarely addressed in formal AI safety research despite being prominent in public discourse.

Current AI safety focuses narrowly on traditional harms: Existing AI safety work concentrates on discrimination, hate speech, harmful content, and malicious uses, but largely ignores risks to human cognition and mental health—creating a gap between academic research priorities and public concerns.

GenAI addiction represents a genuine welfare risk: The paper highlights attachment and dependence on GenAI systems as legitimate concerns worthy of the same attention given to other AI risks, suggesting these psychological impacts deserve formal study and mitigation strategies.

Multiple intervention approaches are needed: The authors propose that addressing cognitive and mental health risks requires coordinated efforts across three areas: improvements to AI safety and alignment work itself, public information campaigns to raise awareness, and regulatory frameworks to protect users.

The discrepancy between research and public concern is quantified: The paper explicitly measures the gap between what AI safety researchers study versus what the public worries about, demonstrating that cognitive and mental health risks are significantly underrepresented in academic literature despite their real-world importance.