| Research Topics: |
| INDEX |
| AI-chat 4 |
| PRIME 33 |
| aging 7 |
| body 1 |
| brain 9 |
| cancer 10 |
| climate 1 |
| cognition 4 |
| covid-19 4 |
| crime 5 |
| culture 2 |
| dental 2 |
| diabetes 3 |
| diagnostics 1 |
| disease 1 |
| drugs 1 |
| economy 1 |
| election 3 |
| energy 1 |
| eyes 1 |
| gender 1 |
| genetics 4 |
| gut 3 |
| heart 5 |
| immune 4 |
| kidneys 1 |
| language 1 |
| metabolism 3 |
| microbiome 1 |
| misc 1 |
| nutrition 4 |
| obesity 1 |
| pain 1 |
| physical 1 |
| politics 8 |
| pollution 6 |
| privacy 1 |
| psychology 5 |
| quantum 3 |
| reproduction 2 |
| rights 2 |
| sleep 1 |
| soc-media |
| space 2 |
| stress 2 |
| tech 1 |
| unk |
| vaccine 4 |
| virus 3 |
| war-confl |
| war-cyber 4 |
|
Scanned: 2595 Evaluated: 284 Selected: 69 |
| AI-CHAT PRIME LANGUAGE PSYCHOLOGY | ||||||||
Mpy29-007 | Prompt Engineering Limitations: Preliminary Evaluation of Large Language Models for Psychotherapy Safety | |||||||
| Quote: | Models frequently validated harmful statements, colluded with hallucinations, minimized symptoms, or used stigmatizing language, including in the newest and largest models. These failures reflect structural limitations such as lack of memory, insufficient contextual reasoning, and training-related biases. Prompt engineering alone is therefore insufficient for safe AI-mediated psychotherapy; clinician-guided fine-tuning, integrated safety mechanisms, and system-level over-sight will be required. This work provides early evidence motivating deeper clinician-led evaluation and safety-oriented model development. | |||||||
| Abstract: | https://www.medrxiv.org/content/10.64898/2026.07.16.26358261v1 | |||||||
| Full Text: | https://www.medrxiv.org/content/10.64898/2026.07.16.26358261v1.full-text | |||||||
| Search Similar: | Brave Mojeek Startpage Dogpile Bing Googl Lycos Utub Yaho | |||||||
| Search Authors: | Nhat Ngo Giang Dao Akane Sano, | |||||||
|
Javilk's Research Links Newsletter for Week 29 Updated 07/25/2026-03:15:16 AM What interests you? What have you found? Tell me! Drop me a note to subscribe or submit something. |