Hello, friends,
In his novel “Cat’s Cradle,” Kurt Vonnegut invents a word, “karass,” to describe teams of people that “do God’s will without ever discovering what they are doing.” He wrote, “If you find your life tangled up with somebody else’s life for no very logical reasons … that person may be a member of your karass.”
Peter Eckersley, who died suddenly at the too-young-age of 43 last week, was in my karass. For more than a decade, he and I often found ourselves walking alongside each other on an intellectual path that led from privacy to artificial intelligence.
We were very different people. He was an Australian computer scientist and lawyer with a wispy beard who biked everywhere, sported colorful socks, and lived in a shared house that he called “the Spaceship.” I am a math major turned journalist who favors preppy non-colorful clothing and lives with my husband and kids in a town house that in no way resembles a spaceship.
But we shared a passion for using computation and automation in our efforts to build a better world. He built tools that supported his advocacy work at the Electronic Frontier Foundation, and I built tools to enhance my journalistic investigations. One time we even built something together. In 2014, we together built a Secure Messenger Scorecard that rated encrypted messaging services. He was one of the first calls I made for advice when Surya Mattu began building his privacy forensic scanner, Blacklight, because Peter had been working to identify commercial tracking for years through EFF’s Privacy Badger project.
Peter was incredibly optimistic and full of joy. While we were both concerned about the problems that technology had wrought—the surveillance, the inequities of access, the abuse of AI—he was always looking for a solution, for a better way. He recognized that to solve a problem, you must first diagnose it. And he was almost always way ahead of the curve with his prognosis and prescriptions.
Consider our first meeting. I was a reporter at The Wall Street Journal who wanted to investigate the emerging commercial market for monetizing personal data. But first I had to learn about the market.
I showed up in January 2010 at a scintillating event on UC Berkeley’s campus called “Exploring Privacy” that was hosted by the Federal Trade Commission. Peter was on a panel with eight other speakers including privacy advocates and tech company executives from Yahoo! and Hewlett-Packard. I was intrigued by his talk about “supercookies,” which he said lurked inside computers and couldn’t be easily removed by users. This sounded like a fascinating story, so I approached him afterward to ask more questions.
In typical Peter fashion, he had already built something that made the concept of undeletable commercial tracking easy for regular people to understand. The day of the event, he published a tool called Panopticlick that let users see exactly how they were being tracked by a type of commercial surveillance known as “browser fingerprinting.”
Showing people how their web browsers were being uniquely identified did more to make tracking real than anything I had been contemplating for my investigation. Inspired, I redoubled my efforts to make my own investigative work more technically revealing. I hired technologist Ashkan Soltani to analyze hundreds of websites and mobile apps to identify tracking technologies as part of our three-year investigative series What They Know. (Soltani is now the executive director of the California Privacy Protection Agency.)
A few years later, when Edward Snowden came forward as a national security surveillance whistleblower, one of his most shocking revelations was how intelligence agencies were exploiting the unencrypted connections of the web to obtain information. Once again, Peter was ahead of the curve. He had already built a tool called HTTPS Everywhere that allowed users to encrypt their traffic and was working with a team to help websites encrypt their connections.
The fruits of that effort, a certificate authority called Let’s Encrypt, launched in 2015. It allowed small websites like my own personal website to encrypt their connections—which had previously been too expensive and technically difficult for small websites to do. By 2020, Let’s Encrypt had issued a billion certificates and contributed to 91 percent of page loads being encrypted in the United States.
Not content to rest on his laurels, Peter soon moved on to tackling problems with artificial intelligence, joining the Partnership on AI in 2018. During these years, our work intersected again, as he investigated how to solve a problem that I had—for the first time—diagnosed before him: the bias in criminal risk scoring algorithms being used in courtrooms across the United States.
I stumbled on the existence of criminal risk scoring algorithms in 2015 and was shocked that they hadn’t been independently evaluated. Working with a team at ProPublica, I obtained thousands of scores that had been assigned to defendants in a Fort Lauderdale, Fla., courtroom and showed that the results were often inaccurate and biased in the way they overpredicted that Black defendants would be arrested in the future.
Peter decided to work toward improving criminal risk scoring systems. His report outlined 10 minimum requirements that should be met for these tools to be used—such as having the tools tested for discrimination and mitigating any detected bias—while also raising the question of whether such tools should be used at all.
Whenever he was in town, Peter would stop by to chat with me about tech policy, and he never left without teaching my kids something new. He taught my son how to make and break ciphers, and he taught my daughter how to crack a Wi-Fi network.
From the day I met him at Berkeley, where he answered all my elementary questions about supercookies, to our most recent Zoom check-in about his new AI institute, he was always extremely generous with his time and enthusiastic about sharing his knowledge.
Peter never stopped building toward a better world. I can think of no better legacy. I’m honored to have had him as part of my karass.
As always, thanks for reading.
Best,
Julia Angwin
The Markup
(Additional Hello World research by Eve Zelickson.)