01
Research
Investigate ethical and societal impacts across the AI lifecycle.
AN AI VISION INITIATIVE · UNITED KINGDOM
Intelligence advances.
Humanity comes first.
Building the foundations for responsible artificial intelligence. A global initiative to research ethics, develop standards and guide the integration of AI into society.
Powered by AI Vision Institute of Technology & AI Vision University
CONCEPT VISUAL · PLANNED FACILITY
✦Human dignity
✦Trustworthy systems
✦Global cooperation
✦Accountable innovation
01 / PURPOSE & VISION
A shared language
for responsible AI.
AI Standards Certification Authority is an initiative within the AI Vision ecosystem. Our ambition is to bring researchers, industry, civil society and policymakers together to develop practical frameworks for AI that serves humanity.
We aim to connect ethical principles with evidence: clear evaluation methods, transparent documentation and meaningful human oversight. Our proposed frameworks will be developed through multidisciplinary research and stakeholder consultation.
02
Standardise
Translate findings into draft requirements, evaluation methods and guidance.
03
Guide
Support informed public policy and responsible organisational adoption.
04
Assure
Develop transparent assessment and certification programmes.
02 / ETHICS & RESEARCH
The questions that
shape our future.
Our proposed research agenda examines how AI affects rights, opportunity and trust — from a classroom assistant to an autonomous robot.
01Human rights & fairness
Investigate discrimination, accessibility and unequal impacts. Evaluate performance across affected groups and provide routes for challenge and remedy.
02Transparency & explainability
Examine disclosure, traceability and explanations that people can understand. Document intended use, limitations and the evidence behind decisions.
03Privacy & data stewardship
Study consent, data provenance, sensitive information and proportionate data use. Connect privacy safeguards with accountability throughout the lifecycle.
04Safety & human oversight
Research robustness, misuse, failure recovery and meaningful human control. Explore escalation, override and safe operating boundaries.
05Generative AI & agents
Study hallucination, synthetic media, intellectual property, tool permissions and autonomous actions. Assess how agents can act safely within delegated authority.
06Society & sustainability
Examine work, education, democratic participation and environmental impacts. Include vulnerable communities in decisions about deployment.
FROM PRINCIPLES TO PRACTICE
Trust is built across
the entire lifecycle.
A proposed evidence pathway for future assessment programmes. Each stage asks a different question; none can be replaced by a single accuracy score.
Define
Purpose, stakeholders and boundaries
Assess
Risks, rights and potential harms
Test
Behaviour, robustness and limitations
Review
Evidence and independent challenge
Monitor
Incidents, changes and improvement
03 / INTERNATIONAL REFERENCE LIBRARY
Grounded in recognised
global frameworks.
A curated starting point for our research and programme design. These references have different roles: management standards, guidance, ethical principles and law.
ISO/IEC 42001:2023
AI management systems
Organisational governance, responsibilities and continual improvement.
Publisher reference ↗ISO/IEC 23894:2023AI risk management guidance
A reference for considering and managing AI-related risks.
Publisher reference ↗ISO/IEC 42005:2025AI system impact assessment
A reference for structured assessment of AI system impacts.
Publisher reference ↗UNESCORecommendation on AI ethics
Human rights and ethical considerations for AI policy and practice.
Publisher reference ↗OECDAI principles
Trustworthy AI, human-centred values and accountable governance.
Publisher reference ↗NIST AI RMFRisk management framework
Voluntary framework organised around Govern, Map, Measure and Manage.
Publisher reference ↗European UnionAI Act
Risk-based legislation; requirements depend on the use and applicable provisions.
Publisher reference ↗Reference does not imply partnership, endorsement or accreditation. ISO standards are copyrighted; this site provides short summaries and links to their publishers.
References reviewed: 13 October 2026
References reviewed: 13 October 2026
OUR PROPOSED STANDARDS / DRAFT v0.1
A practical blueprint
for human-centred AI.
Introducing the proposed ASCA series: ten starting frameworks for research, consultation and future programme design. Open each draft to explore its objective and suggested evidence.
ASCA-001Human dignity & rights
Document the legitimate purpose, affected groups and rights impacts before deployment.
Suggested evidence
Contribute to this draft ↗Impact assessment, consultation record, remedy pathway.
ASCA-002Fairness & inclusion
Evaluate relevant group differences and accessibility; record limitations and mitigation decisions.
Suggested evidence
Contribute to this draft ↗Group-level evaluation, accessibility review, mitigation log.
ASCA-003Transparency & traceability
Identify AI use, document intended capabilities and retain proportionate decision records.
Suggested evidence
Contribute to this draft ↗System card, user disclosures, version history.
ASCA-004Privacy & data stewardship
Record data provenance, legal basis and retention decisions; restrict sensitive-data access.
Suggested evidence
Contribute to this draft ↗Data inventory, access controls, privacy review.
ASCA-005Safety & robustness
Test foreseeable failure modes and misuse; define acceptance criteria for the operating context.
Suggested evidence
Contribute to this draft ↗Test plan, adverse-case results, residual-risk register.
ASCA-006Human oversight & accountability
Assign responsible owners and ensure that authorised people can challenge or stop consequential actions.
Suggested evidence
Contribute to this draft ↗Responsibility matrix, escalation process, override tests.
ASCA-007Generative AI integrity
Evaluate factual reliability, synthetic-content disclosure, source handling and misuse safeguards.
Suggested evidence
Contribute to this draft ↗Grounding tests, disclosure policy, misuse evaluation.
ASCA-008Agentic AI & robotics
Bound delegated authority, minimise tool permissions and test safe fallback and shutdown behaviour.
Suggested evidence
Contribute to this draft ↗Permission inventory, action logs, stop-condition tests.
ASCA-009Public-interest adoption
Use transparent procurement and impact review for AI that affects public services or opportunities.
Suggested evidence
Contribute to this draft ↗Procurement criteria, impact review, public communication.
ASCA-010Sustainability & lifecycle monitoring
Track relevant resource use, incidents and material changes; reassess when deployment conditions change.
Suggested evidence
Contribute to this draft ↗Resource record, incident log, scheduled review plan.
These are original working proposals, not approved international standards or certification criteria. ASCA codes are internal draft identifiers. Detailed scope, requirements, evaluation thresholds, review governance and stakeholder consultation must be completed before adoption.
04 / GOVERNMENTS & PUBLIC INSTITUTIONS
Better policy.
Stronger public trust.
We aim to support governments in understanding AI risks and designing proportionate governance, while keeping human rights and public benefit at the centre.
Our role is research and advisory development. Legal rules are determined by competent public authorities in each jurisdiction.
Policy research & evidence
Comparative analysis, stakeholder consultations and policy briefs to inform decisions.
Responsible public procurement
Proposed guidance on vendor evidence, intended use, oversight, transparency and incident responsibilities.
Regulatory capacity building
Training concepts for public officials, supervisors and multidisciplinary assessment teams.
Sector-specific guidance
Research on education, healthcare, finance, public services and robotics, with context-specific safeguards.
05 / CERTIFICATION DEVELOPMENT
Clear scope.
Credible evidence.
We are developing a portfolio of assessment and learning programmes. Scope, criteria, governance and availability will be published before programmes open.
Professionals
Knowledge & competenceProposed learning and assessment pathways in AI ethics, governance and responsible practice.
Organisations
Governance & readinessProposed evaluation of AI policies, roles, risk processes and documented oversight.
AI systems & robotics
Safety & responsible useProposed system assessments against a defined context, evidence requirements and evaluation criteria.
Programmes are in development. No ISO accreditation, government mandate or internationally recognised certification status is claimed. Future certificates will state their precise scope and limitations.
Certificate verification ↗
A public verification service is planned for issued certificates. It is not yet connected to a live certificate registry.
06 / OUR FUTURE CENTRE
A place to test ideas.
A platform to build trust.
Our vision for a dedicated AI standards and certification centre brings research, testing, education and policy dialogue together.
ILLUSTRATIVE AI-GENERATED RESEARCH VISUAL
Progress with a human purpose.
Our research vision puts human agency, safety and dignity at the centre of intelligent systems.
GLOBAL AI WATCH / NEWS & PERSPECTIVES
Follow the debates.
Understand the implications.
Selected updates from primary sources, alongside our editorial discussion topics. Publication dates are shown. This is a curated snapshot, not a live news feed.
European Commission
9 October 2026 · NEWS
Frontier AI safety under scientific review
The Commission announced a special Scientific Panel meeting addressing frontier AI safety, recent loss-of-control incidents and evaluation methods.
Read original source ↗NIST
19 August 2026 · NEWS
AI-assisted cybersecurity reporting: draft guide
NIST published a draft guide exploring how AI can support CSF 2.0 analysis and reporting. The source lists a public-comment deadline of 15 October 2026.
Read original source ↗UNESCO
20 July 2026 · NEWS
Global learning initiative for AI ethics
UNESCO and LG AI Research launched a free online course translating AI ethics principles into practical skills for professionals and policymakers.
Read original source ↗From the ASCA journal
Why evidence matters more than promises
A discussion prompt: what documentation would let an affected person challenge a consequential AI decision?
ASCA EDITORIAL · DISCUSSION PROMPTWho decides what responsible AI means?
A discussion prompt: how should researchers, communities, industry and public institutions share a voice in standards development?
ASCA EDITORIAL · DISCUSSION PROMPT12-MONTH PROGRAMME / NOV 2026 — OCT 2027
One year of learning.
A continuing global dialogue.
A proposed programme of 12 workshops and 12 discussion forums connecting the draft ASCA series with practical use cases.
24 sessions shown
Human rights & responsible AI foundations
PLANNED · NOT YET CONFIRMEDExpress interest ↗Human rights & responsible AI foundations
PLANNED · NOT YET CONFIRMEDExpress interest ↗AI transparency & explainability
PLANNED · NOT YET CONFIRMEDExpress interest ↗AI transparency & explainability
PLANNED · NOT YET CONFIRMEDExpress interest ↗Fairness, accessibility & inclusion
PLANNED · NOT YET CONFIRMEDExpress interest ↗Fairness, accessibility & inclusion
PLANNED · NOT YET CONFIRMEDExpress interest ↗Privacy & data governance
PLANNED · NOT YET CONFIRMEDExpress interest ↗Privacy & data governance
PLANNED · NOT YET CONFIRMEDExpress interest ↗Robustness, testing & red teaming
PLANNED · NOT YET CONFIRMEDExpress interest ↗Robustness, testing & red teaming
PLANNED · NOT YET CONFIRMEDExpress interest ↗Human oversight & organisational accountability
PLANNED · NOT YET CONFIRMEDExpress interest ↗Human oversight & organisational accountability
PLANNED · NOT YET CONFIRMEDExpress interest ↗Generative AI, misinformation & content integrity
PLANNED · NOT YET CONFIRMEDExpress interest ↗Generative AI, misinformation & content integrity
PLANNED · NOT YET CONFIRMEDExpress interest ↗Autonomous agents & safe robotics
PLANNED · NOT YET CONFIRMEDExpress interest ↗Autonomous agents & safe robotics
PLANNED · NOT YET CONFIRMEDExpress interest ↗AI in education & healthcare
PLANNED · NOT YET CONFIRMEDExpress interest ↗AI in education & healthcare
PLANNED · NOT YET CONFIRMEDExpress interest ↗Public procurement & government guidance
PLANNED · NOT YET CONFIRMEDExpress interest ↗Public procurement & government guidance
PLANNED · NOT YET CONFIRMEDExpress interest ↗Sustainable AI & incident monitoring
PLANNED · NOT YET CONFIRMEDExpress interest ↗Sustainable AI & incident monitoring
PLANNED · NOT YET CONFIRMEDExpress interest ↗Draft standards review & global dialogue
PLANNED · NOT YET CONFIRMEDExpress interest ↗Draft standards review & global dialogue
PLANNED · NOT YET CONFIRMEDExpress interest ↗Planning dates only; events are not confirmed and registration is not open. Proposed online format, 14:00–16:00 Europe/London. Speakers, access arrangements, language support and any fees will be announced separately. Dates and times may change.
QUESTIONS & ANSWERS
Clarity before commitment.
Is this a government regulator?
No. This is an AI Vision initiative developing research, guidance and proposed assessment programmes. Statutory authority rests with the relevant public bodies.
Are the programmes already accredited?
No accredited certification status is claimed. Any future external recognition and programme availability will be published with verifiable details.
How do ISO standards relate to this work?
They are research references. ISO/IEC 42001 concerns organisational AI management; ISO/IEC 23894 provides AI risk guidance. Referencing them does not establish conformity or grant certification rights.
How can organisations participate?
Use the enquiry link to express interest in research, consultation, programme development or policy dialogue. Participation arrangements are subject to review.
What makes an ASCA draft different from ISO?
ASCA drafts are our proposed internal frameworks. ISO standards are developed through ISO’s own processes. The two are not interchangeable, and our codes imply no ISO affiliation.
Can I comment on the proposed standards?
Yes. Express interest by email and identify the draft code, proposed change and supporting evidence. A formal consultation process is planned.
Are the workshop dates confirmed?
They are planning dates for a proposed annual programme. Registration, facilitators, delivery arrangements and any fees are not yet confirmed.
Will workshops provide a certificate?
Attendance or assessment documentation may be considered after programme rules are agreed. Participation alone will not imply professional certification or accreditation.
Can a single test prove that AI is ethical?
No. A responsible evaluation needs a defined context, multiple kinds of evidence, affected-stakeholder input and review over time.
How will conflicts of interest be managed?
Our proposed model calls for disclosure, reviewer independence and separate assessment decisions. Detailed governance and appeals rules are still to be developed.
Can small businesses participate?
The programme design aims to be proportionate to organisational size and risk. Eligibility and practical requirements will be published when programmes open.
Is the planned centre already open?
No operational facility is claimed. The campus image represents the architectural vision for a future centre.
Are news summaries automatically updated?
No. These are dated editorial selections linked to original sources. Updates need editorial review and publication.
Where can I ask about privacy or accessibility?
Contact [email protected] with your question or access needs. The current site uses an email enquiry link; it does not operate an online application database.
BUILD THE FUTURE WITH US
Make responsible AI
a shared endeavour.
We welcome expressions of interest from researchers, public institutions, universities, industry and civil society.
Start a conversation ↗Opens your email application · AI Vision University contact