← Back to Privacy Center

Responsible AI Policy

Effective date: 20 April 2026

1) Purpose

This Responsible AI Policy (the “Policy”) is made available by UniHawk Education Support Services L.L.C, a company incorporated under the laws of the United Arab Emirates, with its registered address at Office 1113, Malik Hamad Suhail Owaida Al Khaili Building, Al Sufouh 2, Dubai, UAE (together, “UniHawk”, “we”, “us”, or “our”) for HawkNest platform, including the website, mobile application, and related features and functionalities (together, the “Platform”).

This Policy sets out the principles, safeguards, and governance measures implemented by UniHawk in relation to the design, deployment, and use of AI within the Platform.

This Policy is intended to align with applicable legal and regulatory frameworks, including the UAE Federal Decree-Law No. 45 of 2021 on the Protection of Personal Data (PDPL), the UAE Child Digital Safety Law (Federal Decree-Law No. 3 of 2024), and other relevant UAE legislation, as well as internationally recognised principles and best practices for responsible and trustworthy Artificial Intelligence (AI).

2) What our AI does

The Platform utilises AI technologies solely to support educational and informational purposes, including:

  • generate guidance, explanations, summaries, and recommendations,
  • help users explore options (universities, courses, pathways),
  • provide conversational assistance.

AI functionalities are designed to be assistive in nature and do not replace human judgment or professional advice.

3) Nature and Limitations of AI Outputs

Users acknowledge and accept that:

  • AI-generated outputs may be inaccurate, incomplete, or outdated;
  • Outputs generated by the Platform (the “Outputs”) are probabilistic in nature and may reflect patterns present in training data; and
  • Outputs should not be relied upon without independent verification from authoritative or official sources.

UniHawk does not warrant the accuracy, completeness, or suitability of AI-generated content.

4) Not for high-stakes decisions

The Platform must not be used as the sole or primary basis for decisions that may materially affect an individual, including but not limited to:

  • academic admissions or placements;
  • scholarships or funding decisions;
  • immigration or visa applications;
  • employment or career outcomes.

For such decisions, we encourage users to consult qualified professionals (such as certified education counsellors, immigration advisors, or legal professionals) and to verify information directly with the relevant institutions or authorities. Users remain solely responsible for all decisions taken based on AI-generated outputs.

5) No foundation model training on student data

UniHawk implements strict controls over the use of personal data in AI systems:

  • Personal data, student-generated content, and user prompts are not used to train, fine-tune, or improve third-party or foundation AI models.
  • Any AI subprocessors (including cloud-based AI providers) are contractually restricted from using such data for independent model training purposes.
  • Service improvement activities are limited to:
    • anonymised and aggregated usage data; and
    • synthetic or fully de-identified datasets for testing and evaluation.

These measures are designed to comply with data protection principles of purpose limitation, data minimisation, and confidentiality under applicable law.

6) Fairness and bias monitoring

We are committed to reducing bias in AI outputs by:

  • reviewing training data and model outputs for patterns of unfairness,
  • monitoring for disparate impact across demographic groups where feasible,
  • iterating on model design and prompt engineering to improve equity,
  • welcoming user feedback on biased or unfair outputs.

Notwithstanding the above, users acknowledge that AI systems may exhibit inherent limitations and unintended bias.

7) Safety and misuse prevention

We aim to reduce harmful use by:

  • applying content and safety controls,
  • monitoring for abuse patterns,
  • enforcing our Acceptable Use Policy.

Where the Platform is accessed by minors, UniHawk shall apply additional age-appropriate safeguards in accordance with applicable laws and internal policies.

8) Privacy and data protection

In accordance with applicable data protection laws, including the UAE Federal Decree-Law No. 45 of 2021 on the Protection of Personal Data, UniHawk:

  • limits Personal Data processing to what is necessary for defined purposes;
  • implements role-based access controls and least-privilege principles;
  • conducts Data Protection Impact Assessments (DPIAs) where AI processing is likely to result in high risk;
  • maintains records of processing activities and vendor/subprocessor assessments; and
  • ensures that cross-border data transfers are subject to appropriate safeguards (including contractual protections or adequacy mechanisms, where applicable).

Where AI functionalities involve automated processing that may significantly affect users, UniHawk implements appropriate safeguards, including, where applicable, the ability for users to request human review or intervention.

9) Transparency and explainability

We aim to make our AI features understandable to users by:

  • clearly labeling AI-generated content where feasible,
  • providing explanations or rationales alongside recommendations,
  • enabling users to ask follow-up questions about guidance provided.

10) Human review and oversight

AI systems within the Platform are subject to human oversight:

  • We encourage users to provide feedback on incorrect or harmful outputs.
  • We may use feedback to improve quality and safety.
  • Users may request human review of any AI-generated output or recommendation that they believe is incorrect, biased, or harmful.
  • We maintain internal processes for reviewing flagged outputs and escalating concerns.

11) Student users (13+)

The Platform is intended for users aged 13 and above. In line with the UAE Child Digital Safety Law, additional safeguards apply to minors:

  • stricter content moderation and safety controls;
  • data minimisation and purpose limitation for all minor-related data;
  • prohibition on targeted advertising, profiling, or commercial exploitation of minor data; and
  • implementation of age-appropriate design and safety measures.

UniHawk does not knowingly permit children under 13 to use the Platform. For more information, please see our Privacy Policy at /legal/privacy-policy.

12) AI safety incident logging and escalation

We maintain procedures for identifying, logging, and escalating AI safety incidents:

  • Logging: All AI safety incidents (harmful outputs, bias reports, data handling anomalies) are logged in an internal incident register with timestamps, severity, and resolution status.
  • Escalation: Critical incidents (e.g., harmful content generated for a minor, data leakage) are escalated within 4 hours to the designated data protection contact and, where applicable, to the school controller.
  • Remediation: Root cause analysis is conducted and corrective measures are documented.
  • Regulatory reporting: To the extent that an AI safety incident constitutes a Personal Data breach, applicable breach notification obligations in accordance with the applicable laws will apply.

Where required under applicable child safety regulations, safeguarding incidents involving minors are handled in accordance with mandatory reporting obligations.

13) Accountability and governance

UniHawk maintains internal governance frameworks for AI oversight, including:

  • We designate internal responsibility for AI safety and ethics review.
  • We periodically review AI systems for compliance with this policy and applicable regulations.
  • We maintain incident response procedures for AI-related safety or privacy issues.

14) Reporting concerns

Users may raise concerns, complaints, or requests in relation to AI use:

15) Updates

UniHawk may update this Policy from time to time to reflect changes in technology, legal requirements, or business practices. The updated version will be published with a revised effective date.