AI Ethics & Governance Integrated System

AI governance assessments, in minutes, not weeks

AEGIS assesses your AI systems the way an expert would. It weighs each factor, shows its reasoning, and grounds every conclusion in the regulation it comes from, so small and growing teams can evaluate compliance with confidence.

EU AI Act Risk Assessment

Regulation (EU) 2024/1689

Describe the system. AEGIS reasons across the prohibited-practice tests (Art. 5), the high-risk schedule (Annex III), and the transparency duties (Art. 50) before reaching a tier.

AI system name Organisation (optional · shown on the PDF report) Sector / domain

// Characteristics — tick all that apply

Makes or materially informs decisions about peoplee.g. approving loans, screening CVs, triaging patients, allocating benefits Processes biometric datafacial recognition, fingerprints, voice, gait — for identification or categorisation Profiles or evaluates peoplescoring, ranking, predicting behaviour, performance or risk of individuals Affects vulnerable groupschildren, elderly, people with disabilities, or those in a position of dependency Generates or manipulates contentchatbot, virtual assistant, synthetic text/image/audio/video (deepfakes) Scores general trustworthiness from social behaviourgeneral-purpose social scoring across unrelated contexts Uses subliminal or exploitative manipulationtechniques that distort behaviour in ways that could cause harm Infers emotions in work or education settingsemotion-recognition of employees or students

Data Protection Impact Assessment

UK GDPR · Article 35

AEGIS applies the Article 35(3) triggers and the ICO's screening criteria, weighs the severity and likelihood of harm, and tells you whether a DPIA is mandatory, advisable, or unlikely to be needed — with reasoning.

Processing activity / system name Organisation (optional · shown on the PDF report)

// Screening — tick all that apply

Automated decision-making with legal or similarly significant effectsArt. 35(3)(a) — decisions made about people with little or no human involvement Large-scale processing of special category or criminal dataArt. 35(3)(b) — health, biometrics, race, religion, sexuality, offences Systematic monitoring of a publicly accessible areaArt. 35(3)(c) — CCTV, tracking, location monitoring at scale Processing data about vulnerable individualschildren, employees, patients, or others with a power imbalance Biometric data used to uniquely identify someonefacial recognition matching, fingerprint identification Large scale — many people, large data volume, or wide geographyaffects a substantial number of individuals Combining or matching datasets from different sourcesdata linking beyond the individual's reasonable expectations Innovative use of new technologyAI, ML, or other novel technology applied to personal data

Model Card Generator

Transparency & accountability

Produce standardised model documentation aligned with EU AI Act technical-documentation expectations and ISO/IEC 42001 transparency principles. Fill what you can — AEGIS structures the rest.

Model name Version Model type Owner / developed by Description & purpose Intended uses (one per line) Out-of-scope uses (one per line) Training data summary Performance & metrics Known limitations (one per line) Bias & fairness considerations (one per line) Human oversight