The Four E-Pillars of AI Adoption: a GTF Government AI Handbook suggestion
The GTF Technologies for Governments Lab (GTF Tampere Lab) has been working, since its meeting in Finland in November 2023, at developing its Government AI Handbook — a useful tool to spark internal conversations within governments at all levels, and to inspire policy decisions and pilot projects.
As a part of the Handbook, GTF has developed the overall thought framework for government AI adoption — the four E-Pillars.
The E-Pillars are not a practical advice, but a general ideation framework that municipalities, departments, and ministries can use in their policymaking.
“This approach is based on what has been identified during the Lab work as the commonplace weakness of most governments: interpretation of AI as a technological, not general, phenomenon; classic ‘roadmap’ methodology poorly adapted to a rapidly evolving technological landscape; cumbersome and too complex interactions with the private sector and own employees.
The Four E-Pillars — namely, Enable, Embrace, Encourage, and Empower — are an element of a possible universal response to these weaknesses. As a general methodology, they can inspire practical solutions on all levels, from municipal to national.”
Successful AI Adoption means govenrment agencies and services are running better, faster, and with less mistakes and noise thanks to AI-based solutions.
-
If the ultimate goal for governments is to create a society and economy that is AI-compatible (and AI-competitive), the first step must be to build a culture of curiosity through education.
Educating different segments of the population about AI should be seen as a precursor to responsible usage and subsequently the ability for users to be able to provide useful feedback upon which existing models can be improved. Amongst non-technical audiences, the growth of AI is daunting and met with fear more than optimism or anticipation. Such attitudes are primarily because of a lack of understanding of the fundamentals of the technology and there is no better way to overcome this than through education. It is important to acknowledge that allowing such mixed feelings and apprehensions to exist with something which is inevitable
In building educational campaigns and programs for AI, governments must place an emphasis on a tailor-made approach. This essentially means acknowledging differences in demographics and indicators such as age, educational background and pre-existing technical expertise in order to provide different audiences with the appropriate resources and material which would facilitate their uptake and use of the tech. It is also important for governments who may choose to run AI educational programs to collect data on what different audiences truly want from the tech. Whilst some audiences may just want to build a basic know-how of generative AI tools, others may want to build a strong understanding of machine learning and neural networks to facilitate their immersion into developing and building models themselves.
At a fundamental level there must be an acknowledgement of the fact that AI itself can only improve through human guidance. The better humans understand the technology, the better they will be able to guide it to improve and this is only possible through education. -
One way or another, AI will affect every part of societal and economic functioning.
Government policy and strategy for a transition of such scale must be built in a manner wherein a wide variety of stakeholders have a seat and say at the table. A government that adopts a multidisciplinary approach is one that is effectively mitigating future risks that will arise from a status-quo wherein there is asymmetric understanding and usage of the technology. It can be said that the most significant challenges arising from AI will not be technical ones but those of education, ethics, equity and economics. It is also important to acknowledge that these above-mentioned challenges will exist across different industries and there will be a fundamental requirement for sectoral expertise to drive how different models are developed and policy is built.
A special emphasis must be placed on involving experts with a social sciences background in discussions about AI policy and strategy. AI is fundamentally replicative of human behaviour, human intelligence & societal dynamics all of which can be understood better through the lens of the social sciences. If the ultimate goal is to build AI that is human-centric, ethical and inclusive, the answer lies in involving thought which understands what humans want, how they think and why they behave the way they do. What we see with AI has perhaps never been seen in history but similarities can be drawn to major historical periods such as the Industrial Revolution and the widespread proliferation of access to the internet. Drawing historical inferences and involving history experts to draw parallels between the socio-economic impacts of such events and the growth of AI could also be useful in how governments respond and build necessary strategy and policy. -
Governing a technology as complex as AI will require a strong comprehension of its capabilities and challenges from a user perspective.
Governments are often accused of sitting in an ivory tower and building policy and regulation at a distance from on-ground realities - AI must be seen as an opportunity to change this. There must be a commitment towards encouraging usage and uptake of AI tools and platforms within government ministries, departments and agencies. It is important to acknowledge that government employees and civil servants are citizens first, part of the broader labour force and therefore must be prepared to deal with the opportunities and challenges of a digital first, AI enabled world.
Perhaps the greatest challenge to widespread adoption and usage of AI will be in building trust in the technology and overcoming inherent apprehensions and fear. To counterbalance this, usage and uptake of the tech within government could serve as a major catalyst. Such an approach would embody the principle of leading by example and inspire trust amongst the masses. Furthermore, the focus should be on encouraging such usage and uptake at the most decentralised levels of government - amongst municipalities and regional governments. Governments at these levels often have a more intimate rapport with their citizens, often as a result of more transparent communication and feedback cycles. Government entities using AI at these levels could also be an extremely effective way in building local champions and encouraging AI driven progress at every level of government. -
In the dynamic world of AI, where technology continuously evolves and reshapes user interactions and needs, the creation of AI policy and strategy must inherently be an iterative process, deeply anchored in user feedback.
This approach recognizes the fast-paced nature of technological change, focusing on understanding how AI usage shifts over time, and how users need to be safeguarded and encouraged to adopt these advancements. Just as AI learns and improves through iterative processes, so too must the policies and strategies that govern it.
For governments, this means embracing the concept of a ‘minimum viable policy.’ This approach involves implementing initial policies, gathering feedback from a broad range of users, evaluating the effectiveness and impact of these policies, and then revising and implementing again. Such a cycle ensures that AI governance remains flexible and responsive to the real-world implications of AI, adapting to both the opportunities and challenges it presents. By prioritizing continuous feedback, governments can create AI regulations and policies that are not only relevant but also resilient to the rapid advancements in the field.
The case for a minimum viable policy is especially pertinent in AI governance due to the technology’s wide-ranging impact across various sectors and its profound influence on societal norms and individual behaviors. This process allows governments to stay ahead of the curve, ensuring that AI is harnessed responsibly and beneficially, while also safeguarding public interest and promoting technological innovation. In essence, by integrating user feedback into the policy-making process, governments can foster a more inclusive, informed, and adaptive approach to AI, positioning themselves as facilitators of progress rather than mere regulators.
The Four E-Pillars is a general approach or a philosophy of Artificial Intelligence adoption for governments. This general approach, born from the GTF Technologies for Governments Lab’s work, is universal and can be applied to any level, from a village to a nation.
However, for obvious reasons, while the approach is the same, its transitions into policies and actions would be completely different depending on a large variety of factors, such as the size of the administrated zone, its economic and demographic context, local culture, etc.
While it is impossible to provide a list of universally functioning conceptual recommendations for this non-universality of precise applications, it is helpful to give at least a glance at possible policy options available for different levels of policymakers across different governance cultures and contexts.
Hence, the GTF Government AI Handbook suggests three policy recommendations and four pilot projects. The GTF team has carefully selected them to either be used as a base for concrete political action or as inspiration for an internal discussion to generate locally adapted solutions.
Each policy recommendation and each pilot project is presented via a list of objectives their application should solve, a general overview of the implementation process, and evaluation criteria.
The two latter sections are given as a reference and are specifically devoid of precise metrics. Even if the Handbook presents the most universally applicable ideas, each must be separately adapted to a given local reality.
The GTF team is happy to provide more insight into the suggested solutions to the leaders interested in exploring their potential in their zone of responsibility.
You can request your copy of the Handbook here.
