Transparency

AI Transparency Notice

How Pekanga uses artificial intelligence to generate career guidance, what data is processed, and the human oversight built into every interaction.

The short version: Pekanga uses AI to generate personalised career pathway suggestions based on a student's subjects, not their identity. No names or personal details are sent to the AI. No student data is retained by the AI provider. All outputs are reviewed by the student before any action is taken. A careers advisor remains central to every meaningful guidance conversation.
What AI does

How AI is used on Pekanga

When a student completes a career exploration session on Pekanga, the platform uses a large language model to generate personalised career guidance. This includes a list of career pathways matched to the student's subject combination, entry routes, AI disruption risk ratings, and hidden gem careers they may not have considered.

Salary data displayed on Pekanga is sourced from the ONS Annual Survey of Hours and Earnings 2025. It is not AI-generated.

What AI does not do

Limits of AI use on this platform

AI model details

Which AI is used and how

AI provider
Anthropic, Inc.
Model
Claude (claude-sonnet series)
Processing location
Anthropic API (USA), zero-retention basis Zero retention
Data sent to AI
Student subjects, qualification type, year group, and (in some modes) career or interest query. No names, email addresses, or persistent identifiers are included in AI prompts.
Data retained by AI
None. Anthropic's API is used on a zero-retention basis. No student data is used to train AI models.
AI disruption risk

The five-tier disruption framework

Every career on Pekanga carries an AI disruption risk rating. Rather than telling students their chosen career is "AI-proof" or vaguely "at risk", Pekanga uses a five-tier framework that describes specifically what AI will do to each role. Honesty about this is a deliberate design choice.

Rating What it means
Resistant AI is unlikely to substantially affect this role. Human judgment, physical dexterity, or emotional intelligence are central and difficult to automate.
Transforms AI will change how this role is performed but will not replace it. Professionals who adapt will become more productive.
Displaces AI will automate a significant proportion of tasks in this role. The role will shrink or evolve substantially over the next decade.
Replaces AI is likely to replace most instances of this role. Students should consider closely adjacent roles or higher-level specialisations.
Creates Demand AI is generating new roles in this area. Demand is growing as a direct result of AI adoption. These are careers that exist because of AI, not despite it.

These ratings are applied by the AI using Pekanga's prompt framework and reviewed periodically. They represent a considered view, not a guarantee. Students should treat them as one input among many.

Human oversight

How people stay in control

Pekanga is designed to support professional careers guidance, not replace it. Human oversight is built into every layer of the platform.

๐Ÿ‘
Student review before action
All AI outputs are reviewed by the student in real time before any action is taken. Nothing is automated or applied without the student seeing it first.
๐Ÿ“‹
Advisor annotation
Careers advisors can view, annotate, and contextualise AI outputs through the appointment notes feature. The advisor's professional judgment sits above the AI output.
๐Ÿ‘
Student feedback
Students can rate any career report with a thumbs up or down. Pekanga reviews this feedback to monitor output quality and identify systematic errors.
๐Ÿ”
Founder review
The AI disruption risk framework and prompt architecture are reviewed by the founder periodically. Known errors are corrected via a maintained override system before they reach students.
Accuracy

What AI can and cannot guarantee

AI-generated career guidance outputs are probabilistic. Career suggestions reflect patterns in the AI model's training data and may not always be current, complete, or accurate. They are a starting point for exploration, not a definitive recommendation. Students are encouraged to discuss outputs with their careers advisor and to explore further using the resources suggested on each career card.

Salary data displayed on Pekanga is drawn from ONS ASHE 2025 (national) and regional data sources, and is updated annually. It is not produced by the AI.

Regulatory context

EU AI Act and UK AI Safety Framework

Career guidance AI is classified as high-risk under Annex III of the EU AI Act (education and vocational training systems that determine access to or outcomes in education). Pekanga acknowledges this classification and voluntarily aligns with its principles as best practice, even though the EU AI Act does not directly apply to UK domestic operations post-Brexit.

Measures maintained in line with this framework include: human oversight built into every interaction, this transparency notice provided to all users, technical documentation of AI model use maintained internally, and data minimisation such that AI prompts contain no names, email addresses, or persistent identifiers.

Questions about AI use

Something not clear?

If you have questions about how AI is used on Pekanga, or if you believe an AI output has caused harm or reflects a systematic error, please contact us. You also have the right to lodge a complaint with the Information Commissioner's Office at ico.org.uk.

This notice is provided in accordance with UK GDPR Articles 13 and 14, the UK AI Safety Framework, and as best practice under the EU AI Act. Last updated: June 2026. Pekanga Ltd, registered in England and Wales.