Ro Encarnación

"Academics can really fill the gap," Encarnación told Penn Engineering in 2024, explaining why she left Capitol Hill for a PhD. "Researchers aren't always reaching out to Congress to offer their expertise in generating policy."

In 2026 she was second author, with Emma Lurie, Sorelle Friedler, and Danaé Metaxa, of the first published study of ChatGPT ads. The team's 91 test accounts collected more than 3,000 ads from 186 advertisers, and lower-income accounts were more likely to see them.

She led a companion AIES 2026 paper on what standard AI benchmarks miss. Running 401 benchmark prompts three times each through ChatGPT's chat interface and OpenAI's API, with and without web search, the team gathered 4,812 responses. Turning on web search cut accuracy by up to 8 percentage points, the same prompt gave inconsistent answers in up to 21% of cases, and the two access routes grounded their answers in different citations.

At FAccT 2026, Encarnación, Christen Hammock Jones, and Metaxa audited how three AI models, including ChatGPT and Gemini, answered realistic questions about abortion law, varying the wording and the U.S. state. Across 54,704 responses checked against state statutes, average accuracy was 78%. The authors estimated that inaccurate answers could expose more than 1 million pregnant people in the United States to misleading information, and found the models often affirmed a question's framing over the correct legal answer.

Her earlier work examined New York City's Local Law 144, which requires bias audits of hiring algorithms. "Auditing the Audits" received an honorable mention for best paper at FAccT 2025. Another study looked at TikTok's AI manga filter and how users themselves exposed its biases, an approach her team called "emergent auditing."

Encarnación grew up in the Bronx, New York, and still remembers the first computer she saw, at a family friend's home. "I would just stand by that room, wanting to use the computer so badly," she recalled. At In-Tech Academy, a combined middle and high school, she learned web design. "Creating websites using computing languages was like magic to me," she said.

Before graduate school she spent a year as a TechCongress fellow for U.S. Representative Yvette D. Clarke. "Going through that experience helped me see how much research is necessary to provide evidence for tech policy," she said. She was one of the founding doctoral students in Penn's Human-Computer Interaction Lab, and she is a GEM fellow and a 2024 recipient of a National Science Foundation Graduate Research Fellowship.