top of page

AI and Data Governance: Insights from the LGAI deep dive

Artificial Intelligence (AI) has transformed the way organizations operate, introducing new tools and challenges. In a recent session focusing on project and data governance, participants gathered to discuss AI's role in public sector projects, its implications for data governance, and the best practices for implementing AI solutions. This blog post provides a comprehensive summary of the key themes and insights from the session, focusing on AI governance, data protection, project management, and ethical considerations.

Setting the Scene: National AI Strategy and Governance

The session began by examining the broader context of AI in the public sector. The national AI strategy sets a vision focusing on economic benefits, while the AI safety agenda prioritizes regulations and guidelines to ensure responsible AI use. Although sector-specific regulation is anticipated, it's crucial to understand the current guidance and principles established by national bodies. The Cabinet Office, for instance, has issued guidelines on the use of generative AI within the civil service, emphasizing safety and ethical considerations.

Key points highlighted include:

  • The emerging role of AI Safety Institutes to manage risk.

  • Sector-specific regulation to address unique challenges.

  • The importance of understanding procurement trends in AI.

Practical Steps to Start an AI Project

Moving from theory to practice, the session discussed best practices for initiating AI projects.

Here are some of the essential steps and recommendations that emerged:

  1. Define Clear Objectives: Establish a clear understanding of the problem that AI aims to solve, ensuring that it aligns with organizational strategy.

  2. Engage Stakeholders: Collaboration is crucial. Engage stakeholders early to gather input and foster buy-in.

  3. Implement Governance: Define roles and responsibilities clearly to maintain control over the AI project.

  4. Communication is Key: Keep all stakeholders informed throughout the project to manage expectations.

  5. Start Small, Scale Fast: Adopt an incremental approach, allowing for flexibility and adjustments along the way.

The importance of communication and stakeholder engagement was underscored, as these elements are critical for managing expectations and achieving successful project outcomes.

Insights from Real-World AI Applications

The session featured case studies illustrating how AI has been used in various settings. These real-world applications highlighted the practical benefits of AI while underscoring the need for governance and data protection.

One example demonstrated the use of generative AI for creating easy-to-read information. Another case study focused on drones for roofing inspections, illustrating how AI can streamline traditionally labor-intensive tasks. A third case study explored AI's role in facilitating translation during emergency situations.

From these examples, several themes emerged:

  • Starting Small: Implementing AI in manageable segments allows for easier management and learning.

  • Solving Real Problems: AI should address specific needs, delivering tangible benefits.

  • Tracking Outcomes: Regularly assess the impact of AI to ensure it's delivering on its promises.

Data Governance and Protection

Data governance was another critical topic discussed in the session. With AI's reliance on data, ensuring data quality, accuracy, and compliance is essential. Here are some of the key takeaways regarding data governance and protection:

  • Data Quality: Ensuring data is complete and accurate is vital. The session noted that up to 80% of time in AI projects is spent on data quality.

  • Data Protection Impact Assessments (DPIAs): These assessments are crucial for identifying potential data risks and ensuring compliance with data protection laws. However, it's important to keep DPIAs updated and ensure they are living documents throughout the project lifecycle.

  • Equalities Impact Assessments (EQIAs): Consider the impact on protected characteristics and ensure AI does not introduce bias.

  • Transparency: Be transparent about how data is used and ensure users understand AI's role in decision-making processes.

  • Accountability: Establish clear accountability for AI decisions and ensure human oversight in critical areas.

The discussion also touched on technical debt, emphasizing the importance of considering the long-term maintenance and validation of AI systems.

Ethical Considerations in AI Projects

Ethical considerations are paramount in AI projects. The session highlighted the following key ethical concerns:

  • Consent and Expectations: Ensure that users understand how their data is used and that they have given proper consent.

  • Transparency and Explainability: Provide clear explanations for AI-driven decisions and ensure users can understand the reasoning behind them.

  • Impact on Staff: Consider the impact of AI on staff, including training and potential role changes.

  • Public Engagement: Engage the public to gather feedback and ensure AI projects meet their needs.

Conclusion and Final Thoughts

The session on project and data governance for AI projects provided valuable insights into the challenges and best practices for implementing AI in the public sector. By focusing on governance, data protection, stakeholder engagement, and ethical considerations, organizations can navigate the complexities of AI and ensure successful outcomes.

It's crucial to keep learning, adapting, and sharing best practices as AI continues to evolve. Collaboration and open communication among stakeholders will be key to achieving responsible and effective AI solutions in the public sector.


bottom of page