All you need to know about the Databricks AI Governance Framework (DAGF) & AI Security Framework 2.0 (DASF)
Warren Atkinson
Cyber Community Connector | Podcast Host | Head of Information & Cyber Security Recruitment | ECS & GTM Team Builder
💻 Latest Episode: https://www.youtube.com/live/rPcDGcZnhis?si=wpn9Ty8a8JI3I9qt
I witness AI’s rapid rise every day, transforming how businesses innovate, operate & engage customers. Generative AI and advanced models are reshaping products, processes and expectations at an overwhelming speed. But with this incredible momentum comes new, complex risks, that many leaders and organisations are still struggling to fully grasp and manage.
I meet with leaders on a daily basis and this is what I’ve learnt. Security, technology & business leaders face a daunting challenge: balancing the incredible promise of AI with the growing threat of misuse, ethical pitfalls, regulatory uncertainty and reputational damage. Many lack the clear visibility and governance structures needed to oversee AI systems effectively, while technology teams wrestle with fragmented model lineage, insufficient access controls, and rapid deployment cycles. Business leaders want to harness AI’s potential but worry about compliance, unintended bias and the possibility of costly mistakes.
This disconnect between AI’s innovation velocity and the organisational visibility required to manage it safely leaves many businesses exposed technically, legally and reputationally. The stakes are high: failure to govern AI properly can lead to data breaches, regulatory fines, lost customer trust, and brand damage that takes years to repair.
That’s why frameworks like the Databricks AI Governance Framework (DAGF) and the AI Security Framework (DASF) are not just helpful… they’re essential. These frameworks provide a strategic, practical blueprint to help leaders align AI initiatives with compliance, ethics, risk management, and security controls, enabling organisations to innovate responsibly and confidently.
Curious to learn which skills are in demand, career pathways and recruitment tips?
The Five Pillars of DAGF: What leaders need to know
DAGF is built on five foundational pillars addressing the full spectrum of AI governance challenges:
Governance Structures: Clear roles and responsibilities to ensure AI accountability at every organisational level.
Legal and Regulatory Compliance: Aligning AI operations with evolving laws and industry standards.
Ethical AI Design: Embedding fairness, transparency, and human oversight into AI systems.
Data and AIOps Infrastructure: Ensuring reliable data lineage, model monitoring, and operational controls.
Security: Implementing lifecycle-aware security controls to protect data and models from evolving threats.
Key takeaways for leaders:
AI risk is multifaceted, encompassing technology, ethics, compliance, and reputation.
Cross-functional collaboration among security, legal, compliance, data, and business teams is crucial.
Transparency and explainability aren’t optional extras; they’re foundations of trust.
Ethical AI must be baked into design and operations, not just documented policies.
Security requires ongoing integration throughout the AI lifecycle.
Learn more and download the whitepaper HERE.
Curious to learn what all the fuss is about?
Deep dive into DASF: The Technical backbone of AI Security
While DAGF outlines governance, the Databricks AI Security Framework (DASF) tackles the detailed technical and operational security risks across AI’s lifecycle. With over 60 AI-specific risks identified and paired with practical controls, DASF secures AI systems from raw data intake to model deployment and inference.
Why DASF is critical:
Lifecycle-Aware Security: AI introduces unique vulnerabilities at every stage, from data collection, model training, and tuning, through to inference and ongoing monitoring. DASF ensures continuous controls like encryption, access management, adversarial testing, and drift detection.
Risk Catalogue & Mitigations: Prioritises risks such as data poisoning, model theft, unauthorised access, and inference attacks, recommending foundational mitigations like authentication, access controls, audit logging, and anomaly detection.
Alignment to Industry Standards: Maps controls to frameworks like NIST, OWASP, ISO, and MITRE, easing integration with existing compliance programmes.
Hybrid Governance & Tooling: Encourages central policies paired with decentralised execution, supported by tools like Databricks Unity Catalog and MLflow for auditability and lineage tracking.
Ethical & Operational Controls: Security extends beyond defence to include fairness audits, bias detection, human oversight, and explainability.
Five actionable insights from DASF:
Embed security throughout the AI lifecycle, don’t treat it as an add-on.
Broaden risk management to include ethical, legal, and reputational factors.
Use hybrid governance models that balance control with developer agility.
Ensure transparency with robust data and model lineage.
Operationalise ethics and fairness via processes and tools, not just policies.
Have you registered your interest?
Why DAGF & DASF matter & how to get started
AI governance and security aren’t optional; they underpin trust, compliance, and sustainable innovation. For leaders in cyber security, infrastructure, and technology go-to-market, DAGF and DASF offer clear, practical frameworks:
Align AI with legal, ethical, and business imperatives.
Manage risk while building trust with customers, regulators, and partners.
Innovate confidently without compromising security or compliance.
Databricks recommends a use-case-driven approach leveraging DASF’s detailed risk catalogue:
Define AI purpose, datasets, model types, and key stakeholders.
Identify top risks and apply foundational mitigations like access controls and audit logging.
Operationalise controls on platforms such as Databricks and Unity Catalog.
Five practical first steps for leaders:
Inventory AI systems and stakeholders to understand exposure.
Prioritise risks with focus on high-impact areas such as public-facing AI endpoints.
Build cross-functional governance teams spanning security, legal, compliance, product, and business functions.
Align with industry standards for compliance and risk.
Define a Minimal Viable Governance Baseline prioritising critical controls.
Have your SECURED your spot?
Join thousands of leading cybersecurity professionals at the International Cyber Expo (30 Sept - 1 Oct 2025, Olympia London) to explore cutting-edge tech from 100+ exhibitors, gain insights from global experts across 3 stages, and network with industry leaders from 85+ countries all under one roof!
Real-world help without breaking the bank: Engaging specialist partners
Effective AI governance doesn’t require costly, large-scale consulting! What organisations need are experienced specialists who’ve been hands-on with frameworks like DAGF and DASF, bringing deep domain expertise and practical insights.
Experts such as Chiru B and the team at Arhasi, Automation Reimagined & Ben Johns at Complyleft, offer global, cross-sector experience and deliver pragmatic, focused support that drives measurable value, quickly and cost-effectively.
Partnering with trusted advisors accelerates your AI governance journey, helping you embed security and ethics without the overheads and delays of traditional consulting.
If you'd like to learn more about how SECURE | CYBER CONNECT & Arhasi, Automation Reimagined can help you, reach out directly to myself Warren Atkinson - we're here to help.
Trust is your greatest competitive advantage
AI isn’t just a technological advance, it’s a profound trust challenge. Customers, partners, regulators & employees all ask: “Can we trust your AI?”
Organisations that embrace frameworks like DAGF and DASF, leverage trusted communities and specialist partners won’t just mitigate risk. They’ll build AI systems that are transparent, accountable and aligned with their values and goals.
This is your moment to lead. Govern AI responsibly and you’ll build trust, scale faster, and shape the future of your industry.
Introducing SECURE | CYBER CONNECT LIVE - Navigating AI Risks, Frameworks & Solutions
We’re proud to present an outstanding panel of experts who are shaping the future of AI and data security. Chiru B, Chief AI Officer at Arhasi, Automation Reimagined, brings over 20 years’ experience in developing secure, governed AI solutions with strong leadership. Arun Pamulapati, Principal Security Engineer at Databricks, is a recognised leader in AI security with more than 25 years’ experience in data protection and compliance. Ramdas Murali, Principal Solutions Architect at Databricks, offers extensive expertise in designing scalable data platforms and delivering enterprise-level architecture. Ben Johns, Cyber Security and Risk Specialist at ComplyLeft, combines two decades of experience in technology and cybersecurity to lead the way in AI trust and safety. Together, they represent the very best in AI innovation and security leadership.
Why This Episode is a Must-Watch & Value You’ll Gain:
If you’re leading AI or security efforts in your organisation, this session is a must-watch, offering practical insights on managing AI risks and compliance while driving innovation. The Databricks AI Security Framework 2.0, reviewed by trusted external partners, mapped to key industry standards, is flexible enough to apply across platforms and sectors, not just Databricks users. This is a great example of global collaboration and the power of community, providing you with the clarity and confidence to lead your AI journey.
Plus, don’t miss the exclusive F*K UPS session on our community platform, where experts share raw, actionable lessons and stories you won’t find anywhere else.
Get a free-copy of the DASF HERE
📺 Watch Session One-off-Two Here: https://www.youtube.com/live/rPcDGcZnhis?si=IClyxZuRba112Kfe
Please show your support by liking, commenting and sharing. Also leaving a comment on what you would like us to address over the coming months!?
Short-form:
We trust you also find value in our Earlier Sessions:
Curious about Web3, Blockchain, DeFi, 5G & IoT? Check out episode Fifty-Four.
Curious about deepfakes, resilience and community power? Check out episode Fifty-Three
Our Podcast Sessions and a range of "Shorts" can be found on YouTube, Spotify, Apple Podcast, X, Instagram, TickTock, Facebook.
✅ Follow, Rate, Subscribe, Like & Share - Simple Search: “Secure Cyber Connect”
SECURE | CYBER CONNECT COMMUNITY – UPDATES
CXO AI Accelerator | Thursday 17th July, London
Join us for an invite-only, intimate workshop will unite senior tech leaders from financial services- CIOs, CTOs, CISOs COO and Director & Heads of, ready to turn AI chaos into a clear, people-first strategy.
Led by a proven transformation expert - Rujuta Singh who’s helped major FS organisations go from idea to execution without losing sight of the humans behind the systems. She brings deep enterprise experience, sharp clarity, and empathetic leadership, with a track record that speaks for itself.
This accelerator workshop comes at No-Cost for our community members and network, because real value should start before the invoice. REGISTER HERE.
Missed the Launch? Catch the F*K UPS Replay - exclusively on our community platform.
Last week we kicked off SECURE | CYBER CONNECT LIVE and the new F*K UPS edition.
Loved the YouTube session? Now dive behind the scenes with F*K UPS - raw, honest talks on personal, professional, health & wellbeing and AI security screw-ups you won’t hear anywhere else! JOIN THE COMMUNIT & WATCH HERE TODAY.
A huge thank you to our partners at Arhasi, Automation Reimagined, Databricks, and Complyleft! Don’t miss the chance to connect with these experts on LinkedIn and YouTube, your support drives the conversation forward. Please watch and follow here: https://youtube.com/@aiwithintegrity?si=ktcNNvhTpAix-cTG
How can we help to address your unique challenges?
We’re more than just a Recruitment Partner, check out our Solutions and Services.
🔗 The SECURE Cyber Connect Directory facilitates Strategic Introductions cross-sector, helping organisations tackle Cultural, Technological & Talent Acquisition challenges, build partnerships, and adapt to regulatory shifts.
A must read:
Reach Out to Warren Atkinson, Justin (Jay) Adamson to explore how we can collaboratively navigate the complexities of AI, Information & Cyber Security to build a safer digital future. We look forward to welcoming you!
Curious to Learn More about the Community, Initiatives & Value provided, click the image below to access our Linktree.
Thanks for sharing, Warren
This offers a much-needed blueprint for managing AI’s complex risks. As AI adoption accelerates, clear governance and security aren’t just best practices-they’re essential to building lasting trust and innovation. When it comes to AI deployment, we encourage our network and community members to reach out to the team directly- Justin (Jay) Adamson & Warren Atkinson. They're here to support you and can help introduce strategies to empower you and your organisation confidently.
AI is moving fast, but governance and security often lag behind. Frameworks like DAGF and DASF provide practical steps to close that gap, helping leaders navigate legal, ethical, and security challenges. For those working with AI in your organisations: what’s been your biggest challenge in AI governance? Drop your thoughts or questions below so we can learn from each other.