Fake names based on ethnicity

[avishwa23]

• Faker version:
• OS: Ubuntu 22.04 LTS

Brief summary of the issue goes here.

I am performing bias evaluation tests for my LLM (language models). While I was able to perform gender bias tests by generating fake male/female data using Faker, I couldnt find a way to generate names based on ethnicity (asian, white, black, etc) to perform ethnicity-based bias tests. Is there a way in Faker to do this?

[fcurella]

There currently is any way to do this, and TBH I don't think we should add such a feature. But you're free to implement your own provider from your own data.

[antik-x]

so?

def test_random_elements():
    from collections import OrderedDict
    faker = Faker(["zh-CN"])
    items = OrderedDict([("亚洲人", 0.25), ("非洲人", 0.25), ("欧洲人", 0.25), ("美洲人", 0.25)])
    print(faker.random_elements(items)[0])

[thijstriemstra]

Interesting question and answer. Looks like you made a typo and meant "There currently isn't any way to do this" @fcurella? In the current AI day and age it's inevitable people will implement this (whether we like it or not).

It's all Fake(r) anyway!

[fcurella]

Beside the fact that it would be considered borderline racist in some cultures, where would you even get the data? I think this would be better served by implementing your own provider.

…

On Wed, Jan 31, 2024 at 5:15 PM Thijs Triemstra ***@***.***> wrote: Interesting question and answer. Looks like you made a typo and meant "There currently isn't any way to do this" @fcurella <https://github.com/fcurella>? — Reply to this email directly, view it on GitHub <#1974 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAAV4B2WIKEQDIX337NCWKTYRLGC3AVCNFSM6AAAAABB3724N2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMRQGE2TGMBQHE> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[github-actions]

This issue is stale because it has been open for 30 days with no activity.