AI & Data Policy

1 How SENSA Uses Artificial Intelligence

1.1 SENSA uses artificial intelligence across three functional areas:

1.2 SENSA does not use AI for:

• Diagnosing or formally assessing special educational needs;
• Making decisions about pupils without human review and sign-off;
• Profiling pupils based on protected characteristics under the Equality Act 2010 in a discriminatory manner;
• Generating content to be presented to parents or local authorities without SENCO review;
• Any automated processing that produces legal or similarly significant effects on a child, without human-in-the-loop oversight.

2 AI Model Transparency

2.1 SENSA currently uses third-party foundation models enhanced with SENSA-specific fine-tuning layers aligned to the SEND Code of Practice 2015, with the aim of using custom models in future. A current list of model providers is published at sensatech.co.uk/ai-models.

2.2 SENSA maintains an AI Model Register documenting: the models in use; their intended function; their input and output data types; and the safeguards applied to each. The Register is available on request to school data controllers, MAT data leads, and DPOs.

2.3 SENSA's need-type adaptation models are reviewed for bias and fairness at least annually against a representative sample of UK pupil demographics. The results will be published in our annual AI Transparency Report.

2.4 SENSA will not knowingly deploy a model that produces systematically different outcomes based on a pupil's ethnicity, gender, socioeconomic status, or English proficiency in a way that is unrelated to the SEN adaptation objective.

3 Data Architecture and Security

3.1 Architecture principles:

3.2 SENSA operates an information security management system (ISMS) aligned to ISO/IEC 27001 principles. Formal certification is targeted before the end of the 2026 calendar year.

3.3 SENSA conducts a Data Protection Impact Assessment (DPIA) for each new feature or significant change to processing operations that may present a high risk to the rights and freedoms of data subjects, in accordance with UK GDPR Article 35.

3.4 In the event of a data breach affecting AI processing pipelines or model inputs/outputs, SENSA's incident response procedure applies in full. See Privacy Policy Section 7 for notification timelines and obligations.

4 Automated Decision-Making and Human Oversight

4.1 In accordance with UK GDPR Article 22, SENSA does not make solely automated decisions about children that produce legal or similarly significant effects. All SENSA Outputs are recommendations subject to human review.

4.2 SENSA Observe produces a ranked watchlist — a prioritisation aid for SENCOs. It is not a diagnosis. The SENCO (or designated teacher) must review all watchlist entries and determine whether further action is warranted. The watchlist cannot be used as evidence of SEN need without professional corroboration.

4.3 SENSA Adapt produces content alternatives. Teachers retain full discretion over which adaptations to use, modify, or discard.

4.4 SENSA Evidence produces draft documentation. All draft content must be reviewed, edited, and approved by the SENCO before submission to any local authority, parent, or external body. SENSA explicitly prohibits the submission of unreviewed AI-generated EHCP evidence.

4.5 SENSA will incorporate explainability features into the Platform to ensure that SENCOs can understand why a pupil has appeared on the watchlist. Need-signal summaries will cite specific observable classroom indicators.

5 Bias, Fairness, and Equality

5.1 SENSA is committed to equitable outcomes for all pupils regardless of their background. We acknowledge that AI systems can encode and amplify existing biases if not carefully designed and monitored.

5.2 Our bias mitigation commitments include:

• Training data review: Foundation models used in SENSA are evaluated for known demographic biases before deployment.
• Output auditing: A random sample of Outputs is reviewed quarterly by our advisory SENCO panel for evidence of systematic variance by demographic group.
• Teacher-facing transparency: Need-signal indicators cite specific, observable classroom behaviours rather than inferential judgements about pupil characteristics.
• Equity testing: SENSA Adapt adaptations are evaluated to ensure that the same learning objective is maintained across all adapted versions — no reduction in academic expectation.

5.3 Schools are encouraged to report any observed output that appears to reflect bias, stigmatise a pupil, or treat a pupil differently based on a protected characteristic. Reports should be sent to: contact@sensatech.co.uk.

6 AI and Safeguarding

6.1 SENSA acknowledges that the Platform operates in a safeguarding-adjacent context. The following safeguarding commitments apply:

• SENSA staff with access to school data will undergo enhanced DBS checks and safeguarding training.
• The Platform includes built-in content filters that detect and flag teacher-uploaded materials containing content inappropriate for a school context.
• Need-signal watchlists indicate potential SEN support needs only — they are not a substitute for the school's statutory safeguarding procedures.
• Where AI processing surfaces content that may indicate a safeguarding concern beyond SEN (e.g. disclosed abuse in pupil writing), the Platform provides a clear prompt to the reviewing teacher to follow the school's safeguarding protocol.
• SENSA will never retain or use any disclosed safeguarding content for model training or product development purposes.

6.2 SENSA's safeguarding lead is contactable at: contact@sensatech.co.uk.

7 Changes to AI Models and Processing

7.1 SENSA will notify schools at least 30 days in advance of any change to:

• The AI models used to process Pupil Data;
• The nature of data inputs to AI models;
• The sub-processors used for AI inference;
• Any feature that materially changes how need-signal generation operates.

7.2 Where a change requires an update to the DPA, SENSA will provide a revised DPA and obtain re-execution before the change takes effect.

7.3 Schools that object to a proposed change may terminate their subscription without penalty by providing written notice within 14 days of the change notification, effective from the date the change would take effect.

8 Compliance Framework

8.1 SENSA operates in compliance with the following legal and regulatory frameworks:

9 Contact and Escalation

9.1 SENSA will acknowledge all written data protection and AI enquiries within 3 business days and respond substantively within one calendar month.