Mind the devolutionary AI gap
Originally published in the Local Government Chronicle [21/01/25]
The AI Opportunities Action Plan as published on Monday is that rare bird. It is a genuinely ambitious grand vision (promising an astonishing end of decade £400bn economic boom through enhancing workplace innovation and productivity) matched by a readily understandable blueprint that points to a technological revolution that could attainably support wholescale public service reform.
Under the aegis of the prime minister’s AI Opportunities Adviser Matt Clifford, the 50 report recommendations fall across three concise report sections to spell out a hopeful path of advance, in admittedly difficult times. In summary, it’s nerdy but not wordy.
As evidence of the Starmer government’s new year reset of its communications strategy, to connect the world of national missions to improving peoples’ lives in tangible and visible ways, potholes and planning were to the fore of the local benefits we can expect.
There may be the usual caveats on offer, that with timely planning permissions granted running at more than 90 per cent, perhaps the issue is more about housing delivery, skills and investment. Or that when it comes to road surfaces it’s not using advance pattern recognition to spot potholes but sufficient resources to fill them with tarmac.
Nevertheless, the corporate challenge for local government will be to deploy AI as a genuine opportunity to modernise the sector. We must consider the timing is that of opportunity, as a moment of ‘kairos’ not mere sequential time ‘chronos’. There is great significance that AI led change is being trumpeted from the centre in the midst of the most epochal change to local government structures and sensibilities in more than half a century. This suggests that, if we are willing to suspend prior cynicism, this time, unlike other e-government or ‘big data’ IT sales bandwagons, AI could be one of the driving forces towards meaningful devolution.
The shift to larger and fewer units of local government and the creation of a new regional/local two tier opens up a nexus where traditionally ‘soft’ public service performance data must mesh with hard economic data to drive place prosperity to inform better local decision making and resource allocation.
The very policy agenda before us presents ideal kindling to spark and fire a local government AI revolution. Especially when we consider the return to regional strategic planning, the shift in established and well-governed combined authorities to single settlements and statutory growth plans, as well as the very need to drive greater cashable frontline efficiencies from reorganisation and its economies of scale ahead of June’s Comprehensive Spending Review.
However, we should take stock before we get lost in the vision of new digital Jerusalems helping engender new earthly Jersualems across the footpath of devolution.
As a reality check, and based on what we know, this process won’t be frictionless or seamless. Training algorithms requires painstaking data preparation, adapting the demands of systems to fit the local context and ongoing vigilance to ensure relevance, human oversight and ongoing investment. So, AI won’t be the panacea. Its success will depend on how well it integrates into local realities on the ground.
The recent parallel with how outsourcing and streamlined shared services were presented as the best answer to rising costs and operational inefficiencies in local government at the start of the millennium is instructive. Certainly, this had its place in a mixed public service economy. But as a ‘silver bullet’ or ‘Easycouncil’ option, it often exposed and widened imbalances in capacity and fatally weakened the coherence of public service delivery.
Similarly, ‘digital by default’ last decade was again a lofty aspiration for accessibility and agility but in many cases poor planning and implementation often led to outcomes the polar opposite of what was promised – entrenching existing inequalities and amazingly creating new inefficiencies.
Indeed, what must be guarded against is a digital devolution divide which would see well-resourced authorities who have the capital, infrastructure and technical expertise storming away with improved efficiency and service delivery, leaving in their wake a long tail of struggling authorities with less strategic capacity and know-how,
This all risks entrenching the ‘Matthew effect’. For those unfamiliar with the Bible, this means ‘to every one who has will more be given; but from him who has not, even what he has will be taken away’. Or in modern lingo, the rich get richer, and the poor get poorer.
Devolutionary AI will, if it is to help all areas succeed, require a level playing field enabling all authorities to harness innovation in collaborative ways that foster a more equitable and efficient future for local government.
AI’s true potential for, among other myriad benefits, anticipating community needs and proactively allocating resources, eating up routine tasks and streamlining workflows, is to free up capable local government staff for higher-value activities and accelerate improvements to local service delivery.
True efficiency will not be achieved by technology alone but through its thoughtful integration. This will require foundational measures such as staff training and improving cross-departmental and pan-service collaboration – steps that can yield meaningful improvement independent of AI.
In this context, AI must be embraced as a complement to human ingenuity in crafting strategies and delivery mechanisms that prioritise collaboration and sustainability to drive meaningful and enduring improvements in place. It mustn’t be a magic technology hammer, knocking already crooked nails further out of place.
Jonathan Werran is chief executive, Localis