Responsible AI Policy

of eFellows Ltd.

eFellows Ltd. (“eFellows”, “we”, “our”) develops, integrates, delivers and uses artificial intelligence (AI) as part of its technology and business services. We are committed to ensuring that AI is used responsibly, safely and transparently, under human oversight and in compliance with applicable law. This policy sets out the principles that guide our approach to AI.

This policy applies to the AI systems that eFellows designs, develops, implements, operates or uses — whether for our own operations or when delivering solutions to clients. It complements our information security and data protection commitments.

Our principles for responsible AI

Human oversight and accountability: AI supports our people. It does not replace their judgement. Outputs generated with AI are reviewed before being relied upon.
Transparency: We are clear about where and how AI is used.
Fairness and non-discrimination: We take care to avoid unfair bias and discriminatory outcomes.
Safety, security and reliability: We design and operate our AI systems to be secure, resilient and dependable.
Privacy and data protection: Personal data in AI contexts is processed in line with the GDPR.
Compliance: Our use and delivery of AI comply with applicable legislation, including Regulation (EU) 2024/1689 (the EU AI Act) and with our contractual obligations.
Risk-based approach: Our controls are proportionate to the risk and potential impact of each AI system. Higher-impact systems are subject to stronger controls and to impact assessment.
Competence and awareness: Our people who develop or use AI are trained and aware of the rules in this policy.

eFellows manages AI through an Artificial Intelligence Management System certified to ISO/IEC 42001:2023, integrated with its quality, information security, privacy and IT service management systems. A designated AI Officer oversees AI governance.

If you have a question or concern relating to our use of AI, you can contact us at office@efellows.bg.

We may update this policy from time to time; any changes take effect once published on this page.