Informed Solutions’ Chief Innovation Officer David Lawton reflects on the current state of AI in the Public Sector and shares his thoughts on the path ahead, outlining how Public Sector leaders can work with industry experts to unlock opportunities for transforming the quality of services for citizens.

Artificial Intelligence (AI) offers massive potential for a public sector, which is caught in the middle of limited investment and a plethora of new economic, societal, and environmental challenges.

We find ourselves at a critical inflection point between ambition and execution for AI in the public sector, and through delivery of award-winning projects with clients, engagement with industry forums and our partnerships with academia, I’ve had the opportunity to reflect on both the challenges and transformative possibilities AI presents.

The Present Challenge: Ambition vs. Execution

The obvious potential for AI in the public sector has led to a surge in pilot and Proof of Concept projects. However, the number of these that have successfully transitioned from experiments to impactful, national-scale deployment remains limited. The Digital Leaders AI Attitudes survey from March 2024 underscores this gap, revealing growing concerns about organisational capability to convert pilots to real-world business value, creating a perception of low return for investment from AI initiatives.

Earlier this year I was invited onto a panel discussion at the Digital Leaders Public Sector Innovation conference, alongside experts from the Department for Works and Pensions, the National Audit Office and Number 10 Downing Street.  During the discussion, I reflected that we’re currently in the foothills of AI adoption in the public sector and there are many parallels with the first wave of digital transformation from a decade ago. In those early stages, people focused on putting paper-based processes online, rather than re-imagining how services and their operating models could be transformed and made ‘digital-native’.

With this experience in mind, it’s easy to draw parallels with a lot of the AI work currently taking place within public sector estates: individual pockets of AI are being deployed as a small part of end-to-end services (e.g. chatbots to route customer calls, Natural Language Processing to search unstructured data, general-purpose productivity tools such as Copilot). That’s not to say that there aren’t significant benefits to be realised by tackling the low-hanging fruit of repetitive manual work, but these advancements only scratch the surface of what is possible.

The real question that emerged during our panel discussion was whether we are truly setting an ambitious enough vision for transforming public services with AI.

The Future Risks: Navigating ‘Rubbish AI’

Looking ahead, the team at Informed is concerned that markets face  a period of ‘rubbish AI’: siloed solutions, quickly deployed to take advantage of open-source data science and cloud-based LLM technologies without a deep understanding of models’ underlying foundations, limitations, and the data on which they were trained.

This not only risks poor service performance but also presents the very real potential for issues relating to privacy, ethics, and bias that could undermine public trust and create  barriers to sustainable transformation.

In parallel, the development of frontier AI technologies will continue their exponential progress, developing and showcasing remarkable technical achievements (for example OpenAI’s Sora text-to-video model). Such developments clearly provide profile and commercial benefits for the organisations that develop them, but they also help to prepare society as a whole for a future world in which these capabilities have a revolutionary impact.

As frontier AI development continues towards the point of Artificial General Intelligence, it will fall on society at large to discuss and decide how and where it’s safe and appropriate to deploy these capabilities. The recent EU AI Act is a great example of developing regulation in this area: clearly setting out situations in which AI systems currently represent unacceptable levels of risk and prohibiting them.

Realising the Opportunity: Informed Solutions’ Approach

For all of the innovation at the frontier, delivering real, on-the-ground change requires the safe and ethical integration of new capabilities into real-world products and services, which in turn requires the transformation of established organisational and service delivery operating models.

At Informed, we’ve focused on the safe, secure, ethical, and effective rollout of nationally-scaled AI programmes like a first-of-kind AI-based decision support platform which is helping NatureScot manage the protection and sustainable development of environmentally sensitive areas across Scotland. We’re thrilled that this work has just been awarded a prestigious Scottish Government Digital Planning Innovation Award based on proven and quantified real-world benefits in efficiency and customer satisfaction. Our teams’ AI and data science capability has also been recognised with a Global Innovation Award from the World Innovation Technology Services Alliance for our work in partnership with the NHS to better understand and learn from patient safety reports across a wide range of care settings.

What’s made these projects so successful has been our approach of embedding data science as an enabler for innovation within multi-disciplinary teams with complementary expertise in user research, service design, security, engineering, and quality assurance.

Just because these new technologies offer a novel approach to delivering services doesn’t negate the value of established disciplines like User Centred Design. Our Communities of Practice are embracing the disruptive and transformative opportunities offered by AI, tackling questions such as: “What does service design look like in a world of pro-active agents and conversational interaction?”, and “What makes for good content design when content is dynamically generated based on a user’s goal?”

Charting the Path Forward

Embracing AI in public services is not just about adopting new technologies; it’s about fundamentally transforming how services are delivered to better meet the needs of citizens. This transformation requires a deep understanding of user needs, a commitment to fairness, security, governance, and a strategic approach to innovation.

By prioritising ethical innovation and user-centric design, we aim to unlock the full potential of AI, ensuring that public services are not only more efficient and responsive but also more aligned with the values and needs of the communities they serve.

In navigating the future of AI in the public sector, we remain focussed on delivering solutions that are not just technologically advanced but also socially acceptable and responsible, and aligned with the broader goals of public service. Through continued education, trust-building, and leadership, we can harness the transformative power of AI to create a brighter, more efficient, and more inclusive future for public services.