The Power of One: How I Found the Strength to Tell the Truth and Why I Blew the Whistle on Facebook

“Lawmakers could use this book for ideas to prioritize aligning AI with human values.”

Some books are seminal not just for their significance but also precision timing.

Whistleblower Frances Haugen’s The Power of One arrives just after AI scientist Geoffrey Hinton publicly left Google to warn of the dangers of his inventions. Together, Haugen and Hinton are like the carbon monoxide detector and fire alarm going off at the same time.

They urge us to shine light on computer code and its stewards, tech companies, which, as Haugen shows, often put children at risk, stoke divisiveness, and threaten democracies.

Humankind’s first encounter with AI was in “recommender systems.” Facebook (Meta) and Google deployed them to target individuals with content that generates clicks—but also happens to rankle and nudge people toward extremes. A girl searches for healthy meals; she is served pro-anorexia sites.

Haugen was an algorithmic product manager who joined Facebook’s Civic Misinformation team to fight lies and hate speech. She quickly discovered Facebook was soft on harm.

In a telling example, Mark Zuckerberg rejected the advice of more than 60 of his experts to turn off the “downstream meaningful social interactions” algorithm, a key driver of “angry clicks.”

While Haugen chronicles such challenges, this is a highly personal book about her formative experiences, including illnesses, leading to her whistleblowing. An intellectually gifted child from Iowa, she aced math competitions and debates. Earning degrees from Olin College of Engineering and Harvard Business School didn’t spare her from patronizing attitudes (and ogling) at male-dominated engineering companies, nor from lingering insecurities.

In her 30s at Facebook, Haugen found herself in a large sandbox of 20-somethings. With the platform’s PR issues mounting, the youngsters—performance metrics and bonuses in mind—were becoming de facto sociologists, political scientists, psychologists, and God-like arbiters of who might suffer.

They trickled out protections (algorithmic tweaks) to countries according to which ones generated the most ad dollars. Consequently, Haugen explains, Facebook in the U.S. differs from that of many other countries, where lies and hate speech circulate even more freely.  

Tech publications lionize lone figures like Zuckerberg, Gates, and Pichai. But Haugen flips the script with her didactic title, The Power of One, underscoring that a single employee—a single woman—successfully challenged a multi-billionaire player.

This book is a treatise on truth, framed in the hope that other tech employees are just as tired of lying. “The future is likely to bring many more Frances Haugens,” she warns tech leaders.

In these pages, tech employees with a conscience can learn how to photograph company documents undetected (using “air gaps”) and collaborate with a trusted journalist, as Haugen did with Wall Street Journal reporter Jeff Horwitz: “He was helping me see the forest, and I was helping him see the trees.”

Many of Haugen’s fresh-faced colleagues pulling the strings at Facebook on democracies, dictatorships, ethnic violence, elections, and tweens (whom Facebook called “herd animals”) quit under pressure. But most simply stepped up to the preposterous challenge, perhaps never realizing they should have training in social sciences. (Haugen notes computer science programs don’t teach that stuff.)

She doesn’t shame them, though. She’s confident that more tech workers will come to her conclusion: “I imagined tossing in bed in 20 years, unable to sleep because I knew I could have acted but didn’t.” Most tech workers are kind and good, she reflects. “People don’t like to have to keep secrets.”

Lest that sound naive, Haugen’s playbook for removing tech’s cloak of secrecy hardly relies on goodwill. Her new career is to build an “ecosystem of accountability,” similar to that around Ralph Nader’s historic car safety campaign (she analyzes the parallels and differences). She aims to prepare “hundreds of thousands of people to hold accountable the information environment of the 21st century.”

Her nonprofit, Beyond the Screen, targets tech where it hurts: training lawyers in tech class-action lawsuits, guiding pension funds to divest, blasting open social media data (accountability has been “starved of sunshine”), and simulating platforms to get under the hood and solve for harms.

Haugen says mandated transparency would give us choices, and many solutions are within reach. Meta might deploy its existing anti-hacker “slowdown code,” for example, to optionally throttle children’s scrolling speed at bedtime, encouraging them to give up and go to sleep, to avoid an Instagram hangover (or skipping school) the next day. Any parent who knows the torment of watching their child’s mind attacked by the mainlining of algorithms would welcome this.

The new AI technology in Chat GPT and other tools is expected to turbocharge the harms Haugen describes, and AI scientists like Hinton say we may be facing an existential threat from machine intelligence. Lawmakers could use this book for ideas to prioritize aligning AI with human values. 

We can only hope that Haugen’s ecosystem of accountability arises in time.