The Centre for Public Impact’s Programme Director Danny Buerkli is adamant that the key element for the successful implementation of AI in public services will be a shift in mind-set. Governments should go from directing policy at people, to shaping it with people
Three points to emphasise:
- If AI is to succeed, the people of the country must be engaged continuously in policymaking that involves AI
- We need to show more attention to the system maintainers. They are just as important as the innovators.
- If AI is to embraced, case studies and engaging stories is what’s needed, not abstract descriptions of the potential usefulness of AI
You will lay out a five-point action plan for governments at the Tallinn Digital Summit 2018. Do you believe that there are enough people in the public sector who have the necessary passion to explore all options within our technological frameworks?
Like in any large organization, you’ll find more motivated people and less motivated people. In my experience, the people working in the public sector that we come across daily in our work at the Centre for Public Impact tend to be extremely driven, passionate and motivated. They believe in our mission to create better outcomes for citizens. They believe that the work they do in government is deeply meaningful and important, and they care a great deal about it. So, I really don’t think that it is a problem to find passionate people in the public sector that are willing to explore the options that AI offers – as long as we build institutions and processes that allow for that.
The best engineers, visionaries and developers usually prefer the more competitive, bureaucracy-free and financially rewarding private sector.
It is true that if we want to harness the power of artificial intelligence – to improve outcomes for governments – we need to think about how we get the necessary expertise and know-how into government. The existing procedures in the process of how we hire and remunerate people need to be tweaked and adapted. There are precedents for how it can happen.
If you look at how it’s been done in the USA and UK, for instance, 18F or the Government Digital Service (GDS), they’ve found ways of bringing incredibly qualified and talented people into government. These are the same private sector experts and top students who wouldn’t have joined without clever hiring processes. The element around pay and progression matters, but more importantly, everyone talks about how millennials (young people in general) crave for meaning in their work. Working in government is a great way to achieve meaningful impact. We need to make sure we explain that to people. If we do that, we will be able to attract qualified top talent.
Another element is that it might not always be desirable for government departments to hire people directly into their structure. Instead, we could have centres of competence or institutions that serve or do work for the government, providing talents with variable interesting career paths.
One of your key points at the Tallinn Digital Summit will be: to make AI a reality, governments’ mind-set must shift from directing policy at people, to shaping it with people. What if the people don’t care? What if they feel they don’t have time to shape the policy – they must worry about their fun and social life first?
Governments too often still prefer the mode of doing policy to people rather than with people. If we’re doing policy to people, they won’t be engaged. How can they be, when they are no more than passive participants?. We need a general mind-set shift. This is particularly important when we use technologies like AI – we need to instil the mind-set that we won’t do this to people, but that we are going to shape AI together. It doesn’t necessarily mean a need for a national vote though or daily interactiins. It could simply mean getting a large enough group of individuals who are affected by this policy into the room to test products and systems with them and ensuring information about how AI is working in government is accessible and available. Only then can they really understand what this means for them. If we do that, I believe people will be happy to engage. It will be a meaningful engagement rather than doing a box ticking consultation exercise or cooking up questions to only hear what we want to hear.
What I’ve learned from my experience is that people tend to care a great deal about the policies and the public services that touch them directly, in a meaningful way.
How do you keep project owners and IT crews motivated during the AI system maintenance period? Even more so with the politicians and civil servants who might feel that the heavy work has already been done?
We are talking about civil servants. But if we generalize for a moment, anyone is motivated by a meaningful shared goal. The motivation becomes a problem if we don’t fully grasp why we’re working on something – whom it is to serve and if we believe the end-user will actively use those technologies that help public services. It’s particularly important for people who don’t work in the so-called frontline, and might not interact with citizens on a daily basis.
I agree that it’s a problem on a larger scale, not just the development and implementing the use of AI – where we tend to focus on innovators. New things. But we tend to forget about , the people who are responsible for keeping things running. I do recognize that we need innovators. But I dare to say that business as usual is just as important, if not more so, than innovating. Yet we struggle to give people doing these jobs enough attention and resources. In that sense, the use of AI in government is no different (in terms of maintenance) than other maintenance issues that we have with buildings, roads and other kinds of infrastructure.
One promise AI has is that it can indeed make a big difference to the quality of *how* government works and the public services we provide. AI isn’t about having a pretty façade, which is aesthetically pleasing. AI is about investing to improve outcomes for citizens.
How can very nuanced AI systems as a debate topic enter the media and political discussion, which is relying more on entertainment, simple visuals, Explain-It-Like-I’m-10 literacy than before? Do you have any ideas on how can journalists keep the AI topic relevant and constant?
The most important thing we can do is to tell specific stories. When we talk about AI, it generally seems hard to understand. AI feels like a a lofty, untouchable, abstract black-box that somehow does things to us. This is a really unhealthy way to think about AI in the long term. We need to talk about specific case studies that involve AI, machine learning, deep learning. Cases where machine learning has helped social workers allocate their time more efficiently, so they can spend their time where it is needed most. Where cities have improved the flow of traffic with help of AI and how AI has reduced the amount of time people spend in traffic jams. How it helps tax officers detect fraud more reliably. That’s how we can keep this relevant and interesting to people.
Then again, there’s no conflict in the story to pull readers in.
The question that must attract the most interest is: “Where could we be if we use this technology appropriately and thoughtfully?” The conflict is between where we are today, and where we could be if we used these technologies in the right way.
Should every country have a minister for digital progress like Sweden and New Zealand do?
There’s no one-size-fits-all solution. For countries like Sweden and New Zealand, having a minister for digital progress may be exactly the thing that is needed. New technologies like AI gain enough attention, time and gravitas. Appointing a minister by itself is not a means to an end.
What seems to be helpful is to have a group in the public sector that has expertise in using new technologies, including AI, and is willing to share both expertise and testing results, so that not every single government office and ministry has to start from scratch.
Do you think it is politically possible to form a small test group of micro-countries who would become the pioneers of AI developing, with Estonia leading the charge? What are the arguments against it?
I know there are a few countries at the cutting edge of using AI and machine learning in government. Estonia is certainly well placed to take a leading role in this. I hope that out of the Tallinn Digital Summit comes a renewed vigour and motivation for countries to start building AI intelligence and machine learning into their public services and government operations. However, we must remind ourselves that even very advanced nations are just beginning to use this technology in earnest. What would make the biggest contribution is if we can learn in the open and share intelligence among countries – and with citizens.