• AI for Good
• Data & Decision Making
• Productivity & Innovation

Written by Chad Bond, Director of Strategy & Innovation, Zazzi

The government’s AI ambition is clear. It wants the public sector to prioritise the adoption of this burgeoning tech. 

The AI Opportunities Action Plan mentioned the need to “push hard on cross-economy AI adoption” and urged the public sector to “rapidly pilot and scale AI products and services.” It’s an understandable demand — AI is seen as a key driver for economic growth and improved public services. 

But there’s a problem. 

Do organisations have the strong digital foundations and high-quality data needed for AI to learn from? And how do legacy technology and manual processes undermine these necessary foundations. 

A report earlier in the year by the Public Accounts Committee warned that out–of–date legacy technology and the poor quality of data and data sharing in the public sector puts AI adoption in the public sector at risk. 

“AI relies on high quality data to learn, but too often government data is of poor quality and locked away in out–of–date legacy IT systems,” it said. AI has the potential to radically change public services but these barriers make it an uphill struggle, it added. 

The report warned there are no quick fixes and calls for remediation funding. 

AI can’t thrive on outdated and disconnected systems — it needs high-quality data on which to learn. 

Legacy systems and under-digitisation: the government’s AI readiness gap 

We know how legacy IT systems and manual processes affect organisations. As we mentioned in a recent blog, they’re costly, inefficient, unreliable, difficult to change, and pose substantial security risks. 

When it comes to data, government departments deal with: 

• Manual data entry and paper forms that lead to duplication and inconsistent records
• Siloed systems and fragmented processes that keep data out of reach and unsharable 
• Poor quality and unstructured data that undermines the ability to analyse and draw meaningful insights for decision-making. 

These aren’t just technical frustrations that hurt service delivery. These data-related issues actively limit what you can do with AI. 

AI needs: 

• Digital processes that can be automated 
• Connected systems that can share information 
• Clean, structured, accessible data that AI can learn from 

For example, for Border Force, we standardised their workflow and eliminated paper-based processes, which enabled data sharing and paved the way for AI integration. Once the foundations were in place, we helped the team explore how AI could further improve operations. 

Until government organisations digitise their services and find an appropriate way to deal with their legacy systems, there’s a danger AI projects will under-deliver, or fail altogether. 

AI is a powerful tool — but only when the foundations are fixed. 

Don’t miss this Digital Leaders AI week webinar: Join Chad and senior experts from the Home Office, the Cabinet Office, and the Department for Business and Trade, as they share insights of what it takes to implement AI effectively and responsibly in government. Register now. 

Originally posted here.

Comments are closed.