Thousands of delegates gathered in London for Innovation 2025, a major event showcasing how government leaders worldwide are leveraging technology and innovation to enhance public services. The conference, held on 25 and 26 March, featured keynotes, main stage presentations, and numerous roundtables discussing cutting-edge topics including artificial intelligence (AI), sustainability, mission-led government, fraud prevention, and economic growth. Sessions are now available on demand, offering valuable insights into the fast-evolving landscape of public sector innovation.

One of the conference’s notable discussions centred on harnessing AI in government, supported by ThoughtWorks. Civil servants from a variety of UK departments and local government explored the practicalities of deploying AI solutions, underscoring a proven "scan, pilot, scale" methodology. This approach involves first researching the right technology, piloting on a small scale with expert teams, and then scaling promising projects. Local governments have demonstrated success in scaling AI initiatives across boroughs, though scaling in central government remains more fragmented due to silos between departments. Participants urged central departments to learn from local government experiences and promote cross-departmental collaboration.

Building or buying AI technologies emerged as another critical theme, with most agreeing that bespoke in-house development is largely unfeasible. Instead, governments must carefully procure tailored solutions that align with specific departmental missions. This necessitates pooling resources across government to meet changing public expectations efficiently. Yet, challenges remain regarding the capacity within many departments to develop AI strategies that address their distinctive needs.

Risk management was also emphasised, particularly the importance of continuously validating that decision-making data sets are representative to prevent model failures. Automated governance plays a crucial role in maintaining compliance as AI models evolve. Trust-building featured prominently—participants highlighted digital identity as a key enabler for broader AI adoption, yet stressed the importance of transparent communication with citizens regarding how AI technologies function and their real-world impacts. Unlike controversial technologies that can be legislated away, AI's integration into public services is inevitable, hence governments must proactively shape its ethical and operational framework. Storytelling about beneficial AI applications within public services, such as in the NHS, was cited as an effective way to build public confidence.

Complementing AI discussions, a roundtable on government sustainability, supported by Signify, focused on making government estates more environmentally friendly. The UK government raised its emissions reduction targets, aiming for an 81% reduction by 2035 en route to net zero by 2050. However, making historic and diverse government buildings more energy efficient remains challenging, especially given the age and complexity of many estates. Incremental improvements like upgrading lighting and installing rooftop solar panels on schools and hospitals were identified as "here and now" opportunities with immediate benefits. Moreover, newer government offices, including second or third headquarters situated around the UK, are often more energy-efficient by design. Interviewees called for sustained cross-government initiatives involving procurement standards and technology deployments that can collectively move the portfolio toward the net zero target.

Another session tackled the broader challenge of “rewiring” the UK government to deliver its five key missions set by Prime Minister Sir Keir Starmer, ranging from economic growth to NHS reform and clean energy leadership. Participants agreed that entrenched bureaucratic cultures and institutional inertia inhibit rapid progress. Civil servants need to overcome a “collective optimism bias” that underestimates the complexity involved and to develop capabilities enabling systemic change rather than incremental tweaks. Leadership was highlighted as critical in guiding departments through the “fogginess” of mission delivery, helping to clarify priorities while balancing ambition and practical execution. The COVID-19 pandemic showcased government’s potential for rapid innovation and risk-taking but was also seen as an exceptional context, not a standard operating model. Lessons from that period, especially in risk tolerance and structured decision-making, remain valuable.

Further exploring AI’s role in wide-ranging government digital innovation, a roundtable supported by Capgemini discussed embedding AI within larger digital transformation efforts. A recurring barrier is the difficulty of scaling innovations beyond pilot stages, with legacy IT systems presenting a significant hurdle. Successes often start with straightforward, user-focused problems, but more ambitious AI opportunities require long-term investment and risk tolerance. Experts called for a shift towards new operating models within government that rethink roles rather than merely automate tasks, noting that all professions will be impacted by AI adoption. Cultivating a “risk-smart” culture, where organizations rapidly prune unpromising pilots, is essential to focus resources on value-generating projects.

Tackling welfare payments fraud and misuse was addressed in another session backed by Visa, which spotlighted digital identity verification and citizen wallets as transformative tools. The UK government estimates welfare fraud and error cost nearly £10 billion annually, with targets to save over £4 billion by 2030. Participants emphasised that trust, transparency, and strong data protection are paramount in digital ID adoption. They advocated for incremental, service-specific digital credentials rather than broad national ID schemes, aiming to reduce citizen burden while improving service delivery and fraud prevention. Biometric authentication can enhance security, especially for vulnerable populations lacking traditional identification. Real-time data sharing and automation could streamline benefits disbursement linked to life events, simplifying public interaction with government. The challenge remains to foster cooperation and data-sharing while respecting privacy, with some envisioning decentralized user-controlled data wallets as a safeguard.

The UK government’s ambition to integrate AI further into public sector productivity was underscored outside the conference as well. For instance, the deployment of an AI tool called Extract, developed by the government’s Incubator for AI, targets planning delays by accelerating the extraction of relevant data from diverse formats, aiming to speed decisions without compromising regulatory rigor. This reflects a broader strategic push demonstrated at London Tech Week 2025, where Prime Minister Sir Keir Starmer announced an unprecedented £1 billion investment in AI compute capacity to position the UK as a global AI leader. This includes scaling computing infrastructure, training millions in AI skills with industry collaboration, and fostering responsible innovation.

Simultaneously, the UK is establishing frameworks for AI safety and governance. The AI Safety Institute, launched with £100 million in funding, works closely with leading AI developers to test and evaluate new models for risks, though it faces challenges such as balancing industry relationships and lacking enforcement powers. This initiative reflects a proactive but cautious stance compared to fluctuating regulatory approaches elsewhere.

Crucially, many conversations emphasise the need for government to overcome siloed innovation, build collaborative ecosystems, and invest in upskilling civil servants to manage AI’s transformational potential responsibly and effectively. The adoption of AI and other emerging technologies comes with both promise and challenges, from ethical considerations, skills development, and legacy systems integration, to maintaining public trust and ensuring equitable service delivery.

Looking ahead, events like the upcoming Digital Government Expo 2025 at London’s ExCeL will further showcase public sector technological advances, offering opportunities for professionals across the UK’s government and public services to collaborate and stay at the forefront of innovation.

Overall, Innovation 2025 and related discussions illustrate an ongoing commitment to leveraging AI and technological innovation within government to improve efficiency, sustainability, and citizen outcomes, while thoughtfully managing the risks and complexities intrinsic to this transformation.

📌 Reference Map:

Source: Noah Wire Services