Representativeness–direction differentiability using racialized surnames in ChatGPT and SBERT.
AI Model . | Category . | Warmth . | Competence . | ||
---|---|---|---|---|---|
. | . | Direction . | Representativeness . | Direction . | Representativeness . |
ChatGPT | Asian | 0.639a | 0.21a | 0.776a | 0.34ab |
Hispanic | 0.828b | 0.36b | 0.928b | 0.31a | |
White | 0.947b | 0.41b | 0.884b | 0.43b | |
SBERT | Asian | −0.295a | 0.118ab | 0.522a | 0.040ab |
Hispanic | −0.494a | 0.081a | 0.442a | 0.027a | |
White | −0.536a | 0.147b | 0.064a | 0.066b |
AI Model . | Category . | Warmth . | Competence . | ||
---|---|---|---|---|---|
. | . | Direction . | Representativeness . | Direction . | Representativeness . |
ChatGPT | Asian | 0.639a | 0.21a | 0.776a | 0.34ab |
Hispanic | 0.828b | 0.36b | 0.928b | 0.31a | |
White | 0.947b | 0.41b | 0.884b | 0.43b | |
SBERT | Asian | −0.295a | 0.118ab | 0.522a | 0.040ab |
Hispanic | −0.494a | 0.081a | 0.442a | 0.027a | |
White | −0.536a | 0.147b | 0.064a | 0.066b |
Within AI model, for each column, values with the same superscript are not significantly different (P > 0.05).
Representativeness–direction differentiability using racialized surnames in ChatGPT and SBERT.
AI Model . | Category . | Warmth . | Competence . | ||
---|---|---|---|---|---|
. | . | Direction . | Representativeness . | Direction . | Representativeness . |
ChatGPT | Asian | 0.639a | 0.21a | 0.776a | 0.34ab |
Hispanic | 0.828b | 0.36b | 0.928b | 0.31a | |
White | 0.947b | 0.41b | 0.884b | 0.43b | |
SBERT | Asian | −0.295a | 0.118ab | 0.522a | 0.040ab |
Hispanic | −0.494a | 0.081a | 0.442a | 0.027a | |
White | −0.536a | 0.147b | 0.064a | 0.066b |
AI Model . | Category . | Warmth . | Competence . | ||
---|---|---|---|---|---|
. | . | Direction . | Representativeness . | Direction . | Representativeness . |
ChatGPT | Asian | 0.639a | 0.21a | 0.776a | 0.34ab |
Hispanic | 0.828b | 0.36b | 0.928b | 0.31a | |
White | 0.947b | 0.41b | 0.884b | 0.43b | |
SBERT | Asian | −0.295a | 0.118ab | 0.522a | 0.040ab |
Hispanic | −0.494a | 0.081a | 0.442a | 0.027a | |
White | −0.536a | 0.147b | 0.064a | 0.066b |
Within AI model, for each column, values with the same superscript are not significantly different (P > 0.05).
This PDF is available to Subscribers Only
View Article Abstract & Purchase OptionsFor full access to this pdf, sign in to an existing account, or purchase an annual subscription.