Research Reveals Bias Toward Additive Advice In Mental Health Support

From "try yoga" to "start journaling," astir intelligence wellness proposal piles connected other tasks. Rarely does it show you to extremity doing thing harmful. New investigation from nan University of Bath and University of Hong Kong shows that this "additive proposal bias" appears everywhere: successful conversations betwixt people, posts connected societal media, and moreover recommendations from AI chatbots. The result? Well-intentioned tips that whitethorn time off group emotion much overwhelmed than helped.

With intelligence wellness problems rising worldwide and services nether strain, friends, family, online communities and AI are often nan first larboard of call. Understanding how we counsel each different could beryllium cardinal to making that support much effective.

A postulation of 8 studies involving hundreds of participants, published successful Communications Psychology, analysed experimental data, real-world Reddit advice, and tested ChatGPT's responses. Participants advised strangers, friends, and themselves connected scenarios involving some harmful habits, for illustration gambling and missing beneficial activities, specified arsenic exercise.

Key findings:

• Additive dominates - Across each context, group suggested adding activities acold much than removing harmful activities.
• Feasibility and use – Doing much was seen to beryllium easier and much beneficial than cutting harmful things out.
• Advice varies by narration – cutting harmful things retired is viewed arsenic easier for our adjacent friends than for ourselves.
• AI mirrors quality bias – ChatGPT gave predominantly additive advice, reflecting patterns successful online societal media.

In theory, bully proposal should equilibrium doing much pinch doing less. But we recovered a accordant tilt towards piling much onto people's plates and moreover AI has learned to do it. While well-meaning, it tin unintentionally make intelligence wellness consciousness for illustration an endless database of chores."

Dr. Tom Barry, Senior Author, Department of Psychology, University of Bath, England

Co-author, Dr. Nadia Adelina from nan Department of Psychology astatine nan University of Hong Kong, Hong Kong said:

"As AI chatbots go a awesome root of intelligence wellness guidance, they consequence amplifying this bias. Building successful prompts to research what group mightiness remove from their lives could make proposal much balanced and little overwhelming."

This investigation was supported by nan Research Promotion Fund of nan Department of Psychology, University of Bath, England.