Read on for highlights from the conversation and listen to the full episode here.
Learning from our Past Experiences
Back in 2008, as the country was experiencing “bad fiscal times” Michael Wilkening – then Under Secretary at CalHHS – met with the agency’s Secretary to strategize about how to make necessary budget reductions work. As they walked through potential cuts, some of the biggest health programs were looking at multiple reductions. The Secretary asked how many families would be hit 2, 3, 4, 5 times with these reductions in services and care.There was no way to know. The data systems were unable to pull that information. There was no common identifier. The best the State team could do at the time was to engage in probabilistic matching to figure out which programs would overlap with one another and how cost reductions might potentially impact families.
What the Open Data Movement Taught Us
In the years that followed, the Open Data Movement spread not only in big cities and states, but globally as well. Government teams engaged data scientists to assist in establishing open data portals, and we all learned about the value of clean data.“If you have bad data, you’re just going to get bad results…you need to have good, clean data sources. You need to have people that understand data science.” – Michael Wilkening
In considering the past 15 years, Jay Nath acknowledges that there has been a big culture shift that required a process of education and mindset changes. Successful government teams recognized that they didn’t “own” the data, they were stewards of this data. There was an open data appetite from not only executives in government, but the community, fed by very clear use cases with measurable value.
“We’re not doing this from an academic standpoint or just ideological standpoint of visibility…there has to be real value to the department, the organization, and to the community writ large,” said Jay Nath. “We did a lot of engagement with the community, with startups, and we saw just a lot of value being created there. And if you step back, there’s been so much economic value and innovation happening with open data.”
Indeed, there are a bounty of data-driven examples – from the human genome project of the 1990s and 2000s to data-informed journalism to restaurant health scores on Yelp – that are both practical and personal.
Government Work is a Team Sport
The open data portals that were central to the Open Data Movement often relied on a “hub and spoke” model with some centralized data and many more spokes of data that jutted outwards in different directions. The challenge with the spoke model is that those off-shoots of data lead to many small data silos – dead ends when it comes to integrating data into more impactful operations.Michael sees this type of model as a very temporary way to go. While there will always be a centralized unit in government for addressing very difficult challenges, governments will be more successful when teams at every level are thinking about how best to leverage data for good.
“I think there are different approaches we need to be thinking about and training up the existing staff, figuring out how to redo some of the duty statements that we have out there, so that we’re hiring the right types of people into into the agencies, just to really modernize the way that government works and delivers services.” – Michael Wilkening
Michael credits great team members for being able to make headway in this space at CalHHS – from legal counsel who helped figure out ways to incorporate innovation into workflows that remained within the guardrails of government to people in non-senior roles. When looking for ways to optimize and improve data operations, he encourages others to think about idea generation in a non-hierarchical way, “as the people who don’t have senior positions, quite often, are the ones who have the best ideas. And so those get filtered out for you, if you keep it in a hierarchical structure.”
A Case for AI in Government
When looking at the decade to come, there’s a lot of potential in unlocking data in legacy systems and being able to ingest that information, and Generative AI may play a role, but there’s a lot to consider.Instead of focusing on the cutting edge of technology, that brings with it untested outcomes, governments should consider opening up the conversation to innovation and “catching up” on technology that has been around for a decade with proven results, like predictive modeling. This is also a great time to outline the risks – especially in areas of health and human services, where people’s lives are at stake – and set clear guidelines of decisions, such as a denial of benefit, that should always be made by someone who is trained and accepts responsibility.
As governments face shortages in qualified candidates for position vacancies, there could also be an opportunity for AI to help. Consider everyday productivity improvements, such as summarizing large documents, or creating a digital log from a call center. While the act itself may only save a few minutes and still require a trained individual to complete, at scale, these time savings can accelerate and augment the staff that are available.
Listen to the full podcast here.
People referenced in this episode:
