How to make government technology better

It’s been able to do this, at least in part, because the city follows an organizational approach similar to one that Jen Pahlka, the founder of Code for America and author of the fabulous book Recoding America, told me about: government technologies are run by in-house product managers who are able to make policy decisions. 

“Some of the most successful legislations are the ones that empower the programs and services where you really have the biggest ability to have tighter feedback loops with the constituents,” said Garces.

Garces told me that the city recently hired the first chief product officer in the country and is building a team of product managers and UX designers to work hand in hand with policymakers. The bottom line is that when people who actually implement policy are able to shape technology, we can get much better results. 

Harlan Weber, a former user experience designer fellow for Massachusetts’s IT department, told me about working on the Common Housing Application for Massachusetts Program (CHAMP) several years ago. He noted that they “went out and did research with tons of people in housing authorities and with government workers who’d have to use the thing.” They then used that feedback, he said, to shape the portal that finally let residents apply for housing benefits in a single streamlined online system.   

Boston has “a lot of inbuilt advantages,” said Weber, also the founder of Code for Boston. “And we’ve worked hard to press those advantages.” 

Massachusetts, he points out, is a highly educated, well-resourced state “that mostly believes that government can be part of the solution and not just part of the problem.” It also helps that Boston is home to a lot of tech companies and tech researchers working in close proximity to the center of government. This has allowed the city to build up an internal talent pool. 

Finally, Boston also has an established culture of prioritizing digital services. The mayor’s office created one of the first government innovation labs in the US, and the city was one of the first to have a chief digital officer and fellows from Code for America. 

All this said, digital services in Massachusetts are far from perfect (and in fact a recent investigation reveals significant problems with CHAMP and affordable housing). As I found in my reporting, there are simply no silver bullets that can fix the government’s broken relationship with technology. It’s just an incredibly thorny problem (which is why this story is part of our new print issue devoted to hard problems!). But it’s critical that governments urgently work to improve digital services—our democracy depends on it. 

I’ve been thinking a lot about something Pahlka said to me about core government services: “If the American public doesn’t see government deliver, I think it’s less that they get driven toward one party or another, and more that they get driven away from government altogether.”

What else I’m reading

• This story from the New Yorker about the inaccuracy of social media posts about the violence between Israel and Hamas is a thought-provoking reflection on the future of our information system, especially during times of crisis. 
• Clearview AI, the face recognition system that scrapes the internet for photos, does not have to pay a $9 million fine to the UK’s Data Protection Agency. The company escaped the massive fee on the grounds that the agency doesn’t have jurisdiction over how foreign law enforcement use British citizens’ data. Clearview is facing several of these fines, which pose an “existential threat” to the company, according to this report from the New York Times’ Kashmir Hill. But this is a sign that perhaps the company will prevail.  
• A 21-year-old computer science student at the University of Nebraska, Lincoln, used AI to identify a word in a charred, 2,000-year-old, tightly wrapped scroll from Pompeii, damaged in the eruption of Mount Vesuvius. The scroll had been incomprehensible, but using a 3D x-ray scanner, the student was able to identify ink patterns and train AI to make out letters that spelled the word for “purple.”

What I learned this week

Google released a policy proposal focused on online safety for kids and teens. It offers several suggestions for legislation, including a risk-based approach for systems to estimate a user’s age and better tools for users to control recommendation algorithms. Perhaps most notable, it recommends a ban on personalized advertising that targets those under 18. Child online safety has been a hot topic in tech policy lately, as I’ve written about, and it’s interesting to get a perspective from Big Tech.