Responsible use of artificial intelligence

AI may be used within WorkplaceHero to support specific, defined activities, including:

• PassportPro and professional development coaching: AI can help users reflect on their professional development by asking questions, identifying themes and suggesting areas for further consideration. It is designed to support reflection, not complete CPD records, reflections or professional evidence on a user's behalf.
• 360 and Organisation 360 reporting: after a feedback exercise has closed, AI may help summarise themes, patterns and recurring observations within eligible feedback. Numerical scores are calculated using defined scoring rules rather than being determined by an AI model. AI-generated narrative is supplementary and should be considered alongside the underlying data.
• Credential and certificate processing: AI may help extract information such as qualification titles, providers, issue dates or expiry dates from uploaded documents. Extracted information must be reviewed or confirmed before it is relied upon as part of a user's record.
• Website and platform assistance: our AI assistant can help users navigate WorkplaceHero, understand our products and services, and find relevant information. It does not replace our team where human support, professional judgement or individual consideration is required.
• Content and service development: AI tools may assist our team with research, drafting, editing, testing, analysis or the development of learning and platform content. AI-assisted work is subject to appropriate human review before publication or operational use.

WorkplaceHero does not use AI to make autonomous decisions about people in areas that could have a significant effect on them.

AI does not independently:

• make recruitment or selection decisions;
• determine pay, promotion, disciplinary action or dismissal;
• decide appraisal or performance ratings;
• determine whether somebody is professionally competent;
• determine whether a learner has achieved a regulated qualification;
• make final assessment decisions on regulated qualification evidence;
• determine safeguarding outcomes or replace safeguarding professionals;
• diagnose health, psychological or behavioural conditions;
• determine an individual's personality or learning style;
• determine compliance status where professional judgement is required; or
• rank individuals for employment or other significant decisions.

Where WorkplaceHero provides scores, progress indicators or other calculated measures, these are based on defined criteria or transparent system rules rather than an AI model making an independent judgement about the individual.

We recognise that the use of AI within education and assessment requires particular care.

AI may support the development and administration of our learning programmes, but it does not replace appropriately qualified tutors, assessors or internal quality assurers.

Assessment decisions for regulated qualifications are made by appropriately authorised people in accordance with the requirements of the relevant qualification and awarding organisation.

Learners may use AI to support their learning where permitted, but assessment evidence must genuinely demonstrate the learner's own knowledge, understanding and competence.

AI-generated work must not be presented as a learner's own assessment evidence where doing so would misrepresent their achievement.

Where there are concerns about the authenticity of assessment evidence, these are considered by people through our assessment and quality assurance processes.

AI does not replace professional safeguarding judgement.

Where an AI-enabled service encounters information that may relate to safeguarding, users should not assume that the technology has identified, assessed or escalated the concern.

Safeguarding concerns should always be reported through the appropriate safeguarding arrangements.

We do not represent AI systems as safeguarding professionals, counsellors, therapists or emergency services.

AI systems are not infallible. They can misunderstand context, produce incomplete information, reflect bias or generate information that appears convincing but is inaccurate.

We design our AI features with these limitations in mind. Where AI-generated information is presented to users, we aim to make its role clear and encourage appropriate human review. AI output should not be treated as fact solely because it was generated by an AI system.

Where accuracy is particularly important, users should refer to the underlying evidence, authoritative source or an appropriately qualified person.

AI systems can reflect biases present within their training data, design or the information provided to them.

We therefore seek to minimise the risk of unfair or discriminatory outcomes by limiting the purposes for which AI can be used, testing relevant features before release and maintaining human oversight.

We do not use AI-generated assumptions about protected characteristics, personality or other inferred personal attributes to make significant decisions about individuals.

If we identify an AI feature that creates an unacceptable risk of unfairness, inaccuracy or harm, we may restrict, modify or withdraw that feature.

Where AI is used to summarise anonymous or aggregated feedback, its purpose is to identify themes within the information provided. It is not designed to identify anonymous respondents.

We do not intentionally use AI to reverse-engineer, infer or reveal the identity of individuals who have provided feedback anonymously.

Where reporting thresholds or aggregation controls are used to protect anonymity, AI does not override those controls.

We apply our normal data protection and security principles when using AI. This means that we aim to:

• use only the information reasonably necessary for the relevant function;
• restrict AI-enabled features to information the user or organisation is authorised to access;
• avoid sending unnecessary personal information to AI providers;
• apply appropriate contractual and security controls to providers processing information on our behalf;
• maintain appropriate records of relevant providers and sub-processors; and
• handle personal data in accordance with our Privacy Notice and applicable data protection law.

We do not use customer or learner data to train publicly available AI models.

Where third-party AI providers process information on our behalf, we select and configure services with appropriate protections concerning the use and retention of that information.

Further information about relevant providers and sub-processors is available through our privacy and security information.

Users should not enter unnecessary confidential, special category or highly sensitive personal information into AI features.

Where a WorkplaceHero AI feature needs access to particular information to perform its intended function, we aim to restrict that access to the minimum information required.

AI features are not intended to provide a mechanism for bypassing existing confidentiality, safeguarding or data protection responsibilities.

Human oversight is fundamental to our use of AI. Depending on the feature, this may include:

• reviewing AI-generated content before publication;
• requiring users to confirm extracted information;
• keeping assessment and qualification decisions with qualified people;
• using defined rules for numerical scoring;
• testing outputs before features are released;
• reviewing unusual or potentially harmful outputs;
• monitoring the continuing suitability of AI-enabled features; and
• withdrawing or changing features where they do not meet our standards.

The presence of AI within a process does not remove WorkplaceHero's responsibility for how that process has been designed or operated.

We aim to be clear when AI plays a meaningful role in a WorkplaceHero feature.

Where appropriate, AI-generated summaries, insights or recommendations will be identified as AI-assisted or AI-generated.

We will not deliberately present an AI-generated judgement as though it were an independent decision made by a qualified professional.

We use AI where we believe it provides a genuine benefit rather than adding it simply because the technology is available.

Where an AI feature is optional, users can choose whether to use it. Core professional development records should not depend on somebody agreeing to use optional generative AI functionality.

We may also place limits on AI features - including usage limits, restricted prompts or predefined workflows - where doing so supports appropriate, safe and proportionate use.

Before introducing a significant AI-enabled feature, we consider:

• its intended purpose;
• whether AI is necessary and proportionate;
• what information the feature needs;
• potential privacy and security implications;
• the consequences of inaccurate output;
• risks of bias or unfairness;
• whether human review is required;
• whether users understand the role AI is playing; and
• whether additional controls are required because of the context in which the feature will be used.

Higher-risk uses require stronger controls than low-risk administrative or reflective uses.

AI technology develops quickly, so responsible use cannot be based on a policy that is written once and forgotten.

We review our AI-enabled features and the controls surrounding them as our platform develops. This includes reviewing prompts and outputs for quality, accuracy, fairness and appropriate tone; reconsidering controls when underlying models or providers materially change; maintaining oversight of AI-assisted learning content; and responding to identified risks or concerns.

We maintain appropriate records relating to the AI systems and providers used within our services.

This Responsible AI Statement is reviewed at least annually and whenever we introduce a significant new use of artificial intelligence.

If you believe an AI-enabled WorkplaceHero feature has produced inaccurate, inappropriate, unfair or potentially harmful information, please tell us.

We would rather investigate a concern than assume an automated system has got something right.

Where appropriate, AI-generated output can be reviewed by a person and errors can be investigated through our normal support, complaints, assessment or safeguarding processes.

Artificial intelligence will continue to change how people learn and work. Our commitment is not to avoid that technology, but to use it deliberately and responsibly.

AI can help us find patterns, reduce administration, support reflection and make professional development more useful. It should not remove human accountability, professional judgement or the right to question how a decision has been reached.

At WorkplaceHero, people remain responsible for decisions about people.