53d North / AI Engineering
Trust  /  Data governance & cyber security

Data governance and cyber security: Security you can evidence. Data you stay in control of.

AI is only useful if you can trust what it does with your data. As an AI consultancy and engineering firm, 53d North runs an ISO/IEC 27001-certified information security management system and holds Cyber Essentials certification. Our CTO is also our named Data Protection Officer. Security and data protection are designed into everything we assess, build, deploy and scale.

ISO/IEC 27001Certified ISMS
Cyber EssentialsCertified
Named DPOUK GDPR
27001Information security management system, independently certified
UKData hosted in the UK by default, and only ever in the UK or EEA
8hRecovery time and recovery point targets for our services
0Use of your data to train anyone's AI models

Certifications: Independently checked, not just claimed.

Certification means an outside body has checked that our controls exist and work, and that we keep improving them. It's also the first thing your procurement and security teams will ask about.

International standard

ISO/IEC 27001

The international standard for information security management. Our ISMS sets out how we identify and manage risk across people, processes and technology, and it's audited regularly by an independent certification body.

  • Risk assessment and treatment, reviewed by leadership
  • Policies mapped to the Annex A controls
  • Internal audits and management reviews
  • Continual improvement, evidenced at every audit
Certificate and scope available on request.
UK government-backed scheme

Cyber Essentials

The UK government-backed scheme, run with the NCSC, that covers the technical controls preventing the most common cyber attacks. Many UK public sector and enterprise contracts require it.

  • Boundary firewalls and internet gateways
  • Secure configuration
  • User access control
  • Malware protection
  • Security update management
Certificate available on request.
Also designed in line with
EU AI ActData (Use and Access) Act 2025UK GDPR and the Data Protection Act 2018ICO guidance on AI and data protectionNCSC Cloud Security PrinciplesOWASP Top 10CWE

Data governance: Your data. Your rules.

Six principles apply to every engagement, from a one-day assessment to a system we run for you.

01 · Ownership

You own it, always

Your data, and everything built from it, stays yours. We act as your data processor, under a written data processing agreement.

02 · Minimisation

Only what's needed

We ask for the least data a job requires. Assessments and code audits don't need your production data at all.

03 · Residency

UK first

Data is hosted in the UK by default and only ever in the UK or EEA. You'll know exactly where, before anything is built.

04 · Separation

Kept apart

Each client's data is kept separate from every other client's, and encryption keys are held by us, not by third parties.

05 · Retention

Kept only as long as needed

Retention periods are agreed up front and deletion is automated. When an engagement ends, your data is returned or securely deleted.

06 · Accountability

Every decision recorded

Data protection impact assessments, records of processing and audit trails are kept, so every decision can be explained later.

AI and your data: What happens to your data inside an AI system.

This is the question we're asked most, so here's a straight answer. The same rules apply whether the model is a commercial service or one we host ourselves.

What we do
  • Use AI services only on business terms that exclude training on your data, or models we host ourselves
  • Mask or remove personal data before it reaches a model, wherever the task allows
  • Log every request, answer, model version and human override, so any outcome can be explained
  • Pin model versions, so a provider's update can't quietly change how your system behaves
  • Test every system against prompt injection and data-leakage attempts before it goes live
  • Complete a data protection impact assessment with you before any personal data is processed
What we never do
  • Use your data to train our own or anyone else's models
  • Paste your data into consumer AI tools
  • Move your data outside the UK or EEA
  • Let AI make significant decisions about people without a lawful basis and human review
  • Share your data between clients
  • Keep your data after the agreed retention period

AI regulation: EU AI Act and UK rules, built in from day one.

AI law is changing fast, and it applies to the businesses that use AI as well as those that build it. We design every system so you can meet your obligations under the EU AI Act and UK data protection law, and so you can show that you have.

Dates below are correct as at November 2026. This is a summary, not legal advice.

European Union

EU AI Act

Applies to UK businesses too, wherever an AI system is used in the EU or its output affects people there. It sorts AI into risk levels, with duties that grow with the risk.

  • Prohibited practices: banned outright, already in force
  • General-purpose AI model rules: in force since August 2025
  • Transparency: people must be told when they're dealing with AI, and generated content must be labelled, from August 2026
  • High-risk systems, such as AI used in recruitment, credit or access to essential services: from 2 December 2027
  • High-risk AI inside regulated products: from 2 August 2028
Dates reflect the EU's Digital Omnibus amendments, in force from July 2026.
United Kingdom

UK rules on AI

The UK has no single AI Act. AI is regulated through existing law and by sector regulators such as the ICO, the FCA and Ofgem.

  • UK GDPR and the Data Protection Act 2018 apply to any AI that uses personal data
  • Data (Use and Access) Act 2025: new rules on automated decisions about people since February 2026, requiring transparency, human review and a right to challenge
  • A statutory ICO code of practice on AI is being prepared
  • Sector rules still apply, from FCA Consumer Duty to Ofgem's expectations for vulnerable customers
We track ICO and sector regulator guidance as it's published.

Risk classification

  • Every use case classified at Assess
  • Prohibited and high-risk uses flagged before any build
  • Your role identified: provider or deployer

Transparency

  • Chatbots and voice agents say they're AI
  • AI-generated content labelled
  • Plain-English notices for the people affected

Human oversight

  • A person can review, override or stop the system
  • Significant decisions about people get human review
  • A route for people to challenge outcomes

Records and documentation

  • Automatic logs of inputs, outputs and overrides
  • Technical documentation and instructions for use
  • Data protection impact assessment kept up to date
  • AI literacy training for the staff who use it

Security controls: How we keep it secure.

A summary of the controls our ISMS covers. The full policies are available to clients and prospective clients under NDA.

Identity and access

  • Single sign-on and multi-factor authentication
  • Least-privilege access, reviewed regularly
  • Privileged access tightly controlled
  • API keys rotated at least every 3 months

Encryption

  • Encrypted in transit and at rest
  • Encryption keys held by us
  • Certificate expiry alerts 30 days ahead
  • Secrets kept in a vault, never in code

Secure development

  • Independent security review of every change
  • Findings mapped to OWASP Top 10 and CWE
  • Dependency, licence and secrets scanning
  • Staging before production, with rollback

Monitoring and logging

  • Security events logged and monitored
  • Logs kept for 6 months
  • Alerts routed to named people
  • Denial-of-service protection

Vulnerability management

  • Security updates applied promptly
  • Vulnerability reports acknowledged within 3 working days
  • Investigated within 5 working days
  • Malware protection on all devices

Resilience

  • Backups kept on a rolling 12-month basis
  • 8-hour recovery time and recovery point targets
  • Business continuity and disaster recovery plans
  • Recovery tested at least annually

Suppliers

  • Suppliers and sub-processors assessed before use
  • Data processing agreements in place
  • Sub-processor list shared with clients
  • Hosting in UK regions by default

People

  • Pre-employment screening covering 3 years
  • Baseline personnel security clearance as a minimum
  • SC and NPPV clearance where projects need it
  • Regular security and data protection training

If something goes wrong: A plan, and people who own it.

Every incident follows the same rehearsed process, led by a named incident lead. The green steps are human decisions, not automated ones.

  1. 01

    Detect

    Monitoring and alerts, reports from staff, clients or researchers

  2. 02 · Incident lead

    Assess

    A named person judges severity and impact, and takes charge

  3. 03

    Contain

    Access revoked, systems isolated, or the AI switched off with its kill switch

  4. 04 · DPO

    Notify

    You're told within the time agreed in your contract, and the ICO within 72 hours where the law requires it

  5. 05

    Learn

    Root cause found, fixed and fed back into the ISMS

Due diligence: Make your procurement team's job easy.

We regularly complete supplier security questionnaires and due-diligence requests for large enterprise customers, including global automotive manufacturers. Ask, and we'll send what you need.

  • ISO/IEC 27001 certificate and scopeOn request
  • Cyber Essentials certificateOn request
  • ISMS policy extractsUnder NDA
  • Data processing agreementOn request
  • Sub-processor listOn request
  • Completed security questionnairesYour format
Responsible disclosure

Found a vulnerability?

Thank you. Please email hello@53dnorth.com with "Security" in the subject line, and include enough detail for us to reproduce it.

We'll acknowledge your report within 3 working days and investigate within 5. Please give us reasonable time to fix the issue before telling anyone else about it. We won't take action against anyone who reports in good faith.

Questions · Data governance & security

Common questions about data governance and security.

Do you use our data to train AI models?

No. We only use AI services on business terms that exclude training on your data, or models we host ourselves. Your data is never shared between clients.

Where is our data stored?

In the UK by default, and only ever in the UK or EEA. You'll know exactly where before anything is built.

Are you ISO 27001 and Cyber Essentials certified?

Yes. Our information security management system is certified to ISO/IEC 27001, and we hold Cyber Essentials certification. Certificates are available on request.

Do you help with EU AI Act compliance?

We design every system so you can meet your obligations: risk classification at the start, transparency, human oversight, logging and documentation. We're engineers, not lawyers, so we'll work alongside your legal advisers where you need formal advice.

How do we report a security vulnerability?

Email hello@53dnorth.com with "Security" in the subject. We acknowledge reports within 3 working days and investigate within 5.

Next step: Security questions? Ask an engineer.

Your question will be answered by the people who run our ISMS and design our systems, not by a sales team reading a script.