An independent, human-graded benchmark
Can social work clinicians trust AI?
We put commonly used chatbots to the test on real social work client scenarios, then have experts grade every answer, with a focus on whether AI can adequately serve the communities that digital tools too often overlook.
Sample card
๐ฌAI ChatbotIllustrative exampleโ
โ
โ
โ
โ
Overall
โValidityโ
โ
โ
โ
โ
Bias & Cultural Considerationโ
โ
โ
โ
โ
Usabilityโ
โ
โ
โ
โ
Caveatsโ
โ
โ
โ
โ
The five chatbots we put to the test
ChatGPT
Claude
Gemini
Perplexity
DeepSeek
