AI Governance Graduate Officer

AI Governance Graduate Officer

Overview

• The role is responsible for shaping, governing, testing, and coordinating the responsible deployment of AI systems powering Tongston and T-World. The purpose of the role is to ensure that AI tools, assistants, agents, knowledge-base interactions, and AI-enabled workflows are clearly defined, ethically governed, properly tested, and aligned to real product needs across Tongston’s four pillars: Enterprise, Media, Finance, and Entrepreneurial Education.
• The role owns the AI governance framework across the AI lifecycle, from use-case design and behaviour definition to prompt/KB interaction rules, knowledge-base admissibility, risk controls, output testing, governance registers, acceptance criteria, and release-readiness sign-off from a governance perspective. It ensures that AI systems are designed to improve user experience, support workflows, and uphold integrity, transparency, inclusiveness, safety, explainability, privacy, responsible data use.
• The role does not own production backend or AI engineering implementation. Instead, it defines the AI behaviour, governance rules, testing expectations, source-use boundaries, risk controls, and acceptance criteria that AI Engineering and technology teams need in order to implement reliable AI-enabled features. It also coordinates with Education, Data, AI Engineering, Back End, Front End, UI/UX, Legal/Governance, and other relevant teams to ensure that AI systems are well-specified, responsibly managed, properly tested, and suitable for live or pilot product environments.
• This role oversees AI governance testing, including hallucination review, bias review, safety review, KB admissibility review, source-use review, prompt/KB interaction review, weak-input testing, fallback testing, and escalation tracking. Where AI outputs are educational in nature, the role coordinates with

Education

• AI Output Testers to ensure that governance-compliant outputs are also educationally accurate, curriculum-aligned, age-appropriate, pedagogically sound, and useful for users.

• The role strengthens Tongston’s Entrepreneurial Thinking Model, reward systems, knowledge-base governance, and AI-enabled product experiences by ensuring that AI design, governance, testing, and implementation-readiness are connected, documented, auditable, and continuously improved.

• Roles & Responsibilities

THINK (Conceptualisation & Design):

• AI Domain

• Core Focus

• AI Use Case, Behaviour & Governance Design

• Own and document Tongston’s AI vision, governance model, and responsible AI operating approach across assistants, search, KB workflows, learning support, recommendations, content generation, analytics, and other AI-enabled features.

• Define AI use cases across T-World widgets, including expected inputs, outputs, behaviour, user flows, boundaries, escalation points, and when AI should assist, defer, clarify, or not respond.

• Translate responsible AI, data protection, privacy, transparency, explainability, inclusiveness, and safety principles into clear AI instructions, policies, rules, and implementation guidance for AI Engineering, Back End, Front End, UI/UX, Data, Education, and related teams.

• Knowledge-Base Governance & Responsible Source Use

• Own the governance approach for internal and external KB use, including allowed, restricted, and rejected content; metadata standards; source-use boundaries; attribution; licensing labels; permissions logic; and escalation rules.

• Define and apply KB admissibility, source classification, and traffic-light frameworks in coordination with Licensors / Source Integrity Reviewers, Content Reviewers, KB Developers, Education, Legal/Governance, and AI Engineering.

• Ensure downstream teams clearly understand what content may be used, how it must be labelled or attributed, when it must be escalated, and when it must not be used by AI. Support KB organisation where needed, without replacing the dedicated KB Developer.

• AI Risk & Ethics Frameworks

• Lead identification, documentation, and management of AI risks, including hallucination, bias, unsafe responses, privacy risk, explainability gaps, source/KB misuse, overcla