Remote Cultural Safety & AI Guardrails Expert

TELUS Digital AI Data Solutions; México; Remote; MXN 464,000 - 1,290,000; Part time

Job summary

TELUS Digital invites experts to join our global AI safety initiative as independent contractors. You will conduct high-stakes adversarial testing sessions, using creative jailbreaking techniques to probe model vulnerabilities and bias risks.

You’ll map regional social, political, and cultural sensitivities to help our models avoid blind spots across markets, synthesize findings into risk reports, and evaluate ethical decision-making in gray-area scenarios to protect users and institutions.

Qualifications

• Education: Bachelor’s degree in Ethics, Cyber Law, Cybersecurity, or related field.
• English proficiency: Minimum B1 (Intermediate) level to document safety findings.
• Language: Native-level fluency in at least one language.
• Context: Deep knowledge of the social and cultural norms of your native language/region.

Responsibilities

• Conduct high-stakes adversarial testing sessions, employing creative 'jailbreaking' techniques to probe model vulnerabilities.
• Map regional social, political, and religious sensitivities to help the model avoid cultural 'blind spots' in diverse global markets.
• Synthesize safety findings into comprehensive risk reports, collaborating with Policy Teams to update model safety guardrails.
• Evaluate the model’s ethical decision‑making in 'gray-area' scenarios, ensuring it prioritizes human safety and institutional values.

Skills

English proficiency (B1)

Native language proficiency

Cultural knowledge

Education

Bachelor’s Degree in Ethics, Cyber Law, Cybersecurity, or related field

Job description

TELUS Digital invites experts to join our global AI safety initiative as independent contractors. You will conduct high-stakes adversarial testing sessions, using creative jailbreaking techniques to probe model vulnerabilities and bias risks.

You’ll map regional social, political, and cultural sensitivities to help our models avoid blind spots across markets, synthesize findings into risk reports, and evaluate ethical decision-making in gray-area scenarios to protect users and institutions.