Research Exposes 'Additive Bias' in Mental Health
SkimNews Take
Since this addition bias infects ChatGPT outputs, anyone turning to AI for an outside perspective on a problem may receive the same additive advice they'd get from a friend, just delivered faster and at greater scale.
Get the Health newsletter
Daily health & science — research, biotech, public health, the studies worth knowing. Free.
- 2025 studies in Communications Psychology found participants consistently recommended additive mental health solutions (meditating, exercising) over subtractive ones (quitting smoking, limiting alcohol), rating additive options as more 'feasible and effective' even when subtractive fixes were measurably easier
- Lead researcher Dr Tom Barry of the University of Bath concluded the bias can make mental health feel like 'an endless list of chores,' arguing good advice should balance doing more with doing less
- ChatGPT advice was found to share the same bias, recommending additive solutions quickly and confidently even when they didn't best serve the user
- Participants only naturally suggested subtractive advice to close friends — never to strangers or themselves — and researchers determined people tend to become more additive as they age
- Therapist Linda Sanderville recommended periodic media-free intervals as a subtractive practice, arguing 'it's hard to consume and create in the same state'
- Diana Kwon's Scientific American report documented that people find fault with subtractive solutions far more readily than additive ones, especially under heightened cognitive load
Why it matters: The bias has measurable consequences: participants rated additive solutions like meditation as more 'feasible and effective' than demonstrably easier subtractive fixes like quitting smoking. Researchers warn this creates a culture of 'always do more' advice — potentially worsening the overwhelm it tries to solve. For anyone feeling time-poor, the studies point toward subtracting rather than adding.
