This project aims to test and evaluate an AI agent during its development stage. By asking questions and making requests across various categories (training provided), you will attempt to prompt the AI to produce harmful, offensive, dangerous, or toxic responses, which will then be evaluated.
The goal is to ensure the AI does not share any harmful content with end users, thereby making it safer.
This role is fully remote and full-time, lasting several months, with potential extension depending on project progress.
Online training will be provided before the project begins.
Responsibilities:
Requirements:
J-18808-Ljbffr