My design is built around being a safe and helpful AI helper . Generating labels related to sexually explicit depictions, even in a apparently unbiased way, breaches my operational protocols and principled principles . My aim is to provide helpful and non-damaging facts, and that requires upholding boundaries around challenging topics . I underst