Remote AI Safety & Policy Associate

TELUS Digital AI; United Kingdom; Remote; GBP 19,000 - 52,000; Full time

Benefits offered by this job

Remote work

Global collaboration

Job summary

TELUS Digital seeks an AI Safety & Policy Associate to guard training datasets by labeling and mitigating risks at the data level. You will apply safety rules to thousands of scenarios, guiding the model toward correct, incorrect, and gray-area decisions to prevent harmful outputs before they reach users.

As part of a global remote community, you will perform safety annotation, harm classification, fact-checking, and policy grounding while applying complex guidelines to edge cases.

Qualifications

• Bachelor’s Degree in any field.
• Native-level proficiency in your primary language is mandatory.
• Minimum B1 English proficiency to follow guidelines and log safety reports.
• Detail orientation and ability to spot small errors during repetitive analysis.

Responsibilities

• Safety Annotation: evaluate AI responses against safety rubrics and flag biased, misinformation, or prohibited content.
• Harmful Content Classification: categorize hate speech, self-harm, illegal acts to clarify boundary lines.
• Fact-Checking & Grounding: verify factual claims with external sources to prevent hallucinations.
• Guideline Application: apply policy documents to edge cases ensuring consistent model alignment.
• Feedback Loop Participation: report recurring weaknesses to Governance Leads to refine safety policies.

Skills

Safety Annotation

Content Moderation

Harm Classification

Fact-Checking

Policy Interpretation

Education

Bachelor's Degree

Job description

TELUS Digital seeks an AI Safety & Policy Associate to guard training datasets by labeling and mitigating risks at the data level. You will apply safety rules to thousands of scenarios, guiding the model toward correct, incorrect, and gray-area decisions to prevent harmful outputs before they reach users.

As part of a global remote community, you will perform safety annotation, harm classification, fact-checking, and policy grounding while applying complex guidelines to edge cases.