Solving civic tech challenges at Code for America Summit

Profile picture for user Matt Jukes By Matt Jukes June 4, 2019
Summary:
Matt Jukes shares what he learned at the recent Code for America Summit, which is one of the must attend events for those working in civic tech.

Image of Bay Bridge in San Francisco

The Code for America Summit is perhaps the biggest civic tech/digital government conference happening anywhere in the world these days. 1300 attendees, passionate about improving public service through technology and service design descended on Oakland for three days last week.

This was the first time I had attended the Summit and while I knew various other attendees, I was attending solo, so it could at times feel quite overwhelming. I had a similar feeling attending SXSW in Austin a few years ago - the difference here though were the kinds of people attending.

There were representatives from pretty much every State across the US as well as further afield (Canada, Mexico, Argentina, Japan and the UK were well represented) with elected officials, civil servants, consultants, civic hackers and a 100 other job roles involved.

With any conference of this scale there is inevitably a significant FOMO factor and I was regularly forced to trade off one session for another based on little more than a short abstract and a gut feeling. 

I was lucky though (or maybe the conference was just that well planned) to attend a series of intriguing, interesting and occasionally inspiring talks.

I am not sure whether it was reassuring or worrying how many of the common themes from similar convenings in the UK kept coming up. Time and again the topics of procurement and hiring technologists & designers came up. There were multiple sessions and workshops on those topics as well. If anything it felt like the problems faced by our US cousins in these areas are tougher than ours (albeit with very similar DNA) and it is no picnic in Canada either.

Jennifer Anastasoff and Jenny Smith - both formerly in leadership roles in the 'people operations' team at the US Digital Service - presented some compelling ideas on how to improve hiring of technologists (and designers, data scientists, product people and the like) into public service based on their experience of bringing 350 people into USDS in just over two years. They have also written a useful pamphlet called 'Mobilizing Tech Talent' which outlines their approach and thinking. It is well worth a read.

Retroactively clearing convictions

The most inspiring slot of the event was the one-two punch on Friday morning of the interview with the Ear Hustle team and then the session about automatically retroactively clearing convictions for marijuana related crimes in partnership with the San Francisco District Attorney after Proposition 64 legalising cannabis was passed.

Ear Hustle is a podcast that tell stories from inside the walls of San Quentin prison. It is presented by Nigel Poor, an artist and activist who was volunteering at the prison and Earlonne Woods who until his sentence was commuted in November 2018 was a prisoner who 25 years to life. They were both in attendance and interviewed on the main stage - it was inspiring and a little heart breaking to hear the stories they had to tell - the 3 strikes law in the US is horribly draconian and the numbers they were talking about were staggering.

Also Earlonne probably had the quote of the conference - "Being free ain't free."

The story of Code for America's work with clearing records was really a tale of two halves.

The evolution of the product itself is pretty fascinating - starting as a web based service that simplified the complex and paperwork heavy process to apply to have your record cleared when appropriate and sought to match applicants up with the legal representation they needed to see things through. It was a success from a product point of view but on closer investigation (and a number of iterations) they realised they weren't really moving the needle on getting significant numbers to take advantage of the service. There were obstacles across the piece - not least that not every California County (responsible for processing the applications) had assigned lawyers and the ones that did were over worked and under resourced.

Their new focus is a system that automatically clears the records of any and all individuals who qualify for it. The system reads a persons criminal record, then maps the data to see if they are eligible to have their record cleared for any convictions and then it completes the appropriate forms to be filed with the court without human intervention.

The public trial of this system was with the San Francisco District Attorney's Office who had committed to clearing cannabis related convictions after the law changed in the City. They had started trying to proactively resolve the backlog manually but with the help of Code for America they were able to automate things and were able to -

  • Clear 9,361 convictions dating back to 1975
  • 1,336 of who now have zero felonies on their record (so important with the 3 Strikes rule!)
  • 729 people now have no criminal record at all

Before this only 23 of people who qualified for the record clearance had even applied! 

Given a criminal record in the US makes getting a job incredibly difficult and things like social housing really hard this has an enormously positive social effect.

This partnership between civic tech and public servants with a clear mission is a real case study in getting this stuff right...and it has been noticed. States and cities across the US have been in touch with the DAs office and Code for America to see if they can re-use the approach.

Other insights

Related to the change of the law in California was the need for the City & County of San Francisco to create a licensing system for potential shop owners. The responsibility for this fell to the Digital Services team (led by former Futuregov-er Carrie Bishop) and they shared their 'Trials and Tribulations of Creating Online Permitting in the Cannabis Industry'. There were a few main takeaways from the session - working with big vendors is never as straightforward as you'd hope - even if their HQ is just down the street. Including potential users in the process might be good practice but it can also have unexpected legal consequences when competitors complain that those involved in the research had an unfair advantage. Finally building a service when the legislation is still being written is like playing football (either version) when the goalposts keep moving.

My final session was - unsurprisingly given my time at the Office for National Statistics - about the work of the US Census Bureau on the upcoming 2020 Census.

The US Census faces a lot of challenges - funding has been cut, controversial questions are being challenged in Court, trust in Government is at an all time low and this is also the first 'online first' Census which brings a new set of challenges.

A lot of the conversation was focused on just how central to the US democratic process (the amount of seats a district gets in Congress are decided based on Census population figures) and the challenges of reaching 'hard to count' communities. In the current climate there is a lack of trust in how the data collected will be used and that it will remain anonymous (whether through State actions or from outside attacks) and much of their communications are focused on changing this perception.

They are investing heavily in chatbots and text message based tooling in an effort to broaden participation and take some of the admin burden from local civil servants (the first call for many US citizens on receiving any communication from any level of Government is apparently their city level public services - who often have little more information than the recipients of the communications - so the hope is these tools will lessen the burden and support multiple constituencies.

Like here in the UK demand for Census data is high and the Census Bureau teams are rightly proud of their work preparing to make more of the data available via a range of tools and formats. The mix of data visualisations, bulk download and API tools is similar to the work of ONS here - though, forgive me a moment of pride, I would suggest the UK is ahead in most of those areas.

I was also lucky enough to attend sessions about the great work of the Canadian Digital Service, the Government as a Platform programme in Argentina, the work of Code for America on making it easier for Californians to apply for food stamps, the need for new thinking on the topic of ethics and responsibilities in tech culture, the work of 18F supporting the US Forestry Service and a 'state of the digital nation' address from Matt Cutts who leads the US Digital Service (and was formerly a somewhat well known Google-er). Phew.

My take

All in all it was a great event - well organised, friendly and with a constant stream of quality talks, panels and workshops. I would recommend it to anybody with an interest in civic tech or digital government - while it is obviously US focused so many of the challenges and approaches are universal.

Next year the Summit moves to Washington, DC for the first time (and will be earlier in the year - taking place in March). I'll be returning - hopefully I'll see a few more UK faces there.