PeopleofSuperintelligence

Researcher

Daphne Ippolito

Daphne Ippolito is an assistant professor in Carnegie Mellon University's Language Technologies Institute and a senior research scientist at Google DeepMind who studies language generation, memorization, and the privacy and security of language models. During her computer science doctorate at the University of Pennsylvania, which she completed in 2022 while working part-time at Google Brain, she researched story infilling, machine-generated text detection with tools such as RoFT, and training-data deduplication. She also coauthored a November 2022 study of professional writers using the Wordcraft AI writing assistant and a November 2023 paper extracting gigabytes of training data from models including ChatGPT. At Carnegie Mellon, she presented research on training-data poisoning at CyLab in February 2025 and co-teaches the graduate course Large Language Model Applications in Spring 2026.

Badges

Timeline

  1. 2019

    Hierarchical story infilling

    In 2019, she coauthors “Unsupervised Hierarchical Story Infilling,” a two-stage method that selects rare words and predicts words to fill missing spans in a story. It aims to balance fluency and coherence with diversity in interactive tools for creative writing.

  2. 2020

    Detecting machine-generated text and RoFT

    In 2020, she is first author of an ACL study of decoding and sampling methods for generated text, finding that methods tuned to fool people introduce statistical abnormalities that automatic systems detect. At EMNLP 2020, she coauthors RoFT, an evaluation tool for detecting the boundary where a human-written passage transitions to machine-generated text.

  3. 2021–2022

    Google Brain and a Penn PhD

    In 2021, as a final-year doctoral student at the University of Pennsylvania co-advised by Chris Callison-Burch and Douglas Eck, she participates in the MIT EECS Rising Stars workshop and begins working part-time as a research scientist at Google Brain. In 2022, she completes a Ph.D. in Computer Science at the University of Pennsylvania and coauthors an ACL paper showing that deduplicating training data cuts memorized text emission tenfold.

  4. November 2022

    Wordcraft with professional writers

    In November 2022, she coauthors a study in which 13 professional, published writers use Wordcraft, a text editor with built-in AI-powered writing assistance tools, to craft stories.

  5. July 2023

    Teaching generative AI at CMU

    In July 2023, she teaches a tutorial on building generative AI systems at Carnegie Mellon. She is an assistant professor in the Language Technologies Institute and a research scientist at Google.

  6. November 2023

    Extracting training data from ChatGPT

    In November 2023, she coauthors “Scalable Extraction of Training Data from (Production) Language Models.” The paper shows that an adversary can extract gigabytes of training data from open-source language models such as Pythia and GPT-Neo, semi-open models such as LLaMA and Falcon, and closed models including ChatGPT.

  7. February 2025

    Training-data poisoning at CyLab

    At a CyLab seminar in February 2025, she presents research on training-data poisoning. The work investigates whether poisoning portions of the internet allows an adversary to insert backdoors or alter the behavior of language models trained on that data.

  8. As of October 2026

    CMU and Google DeepMind

    In Spring 2026, she co-teaches Carnegie Mellon's graduate course Large Language Model Applications with Fernando Diaz. As of October 2026, she is an assistant professor in the Language Technologies Institute and a senior research scientist at Google DeepMind. Her research examines privacy and security in language generation and evaluation of language models.

Career

  1. NowCarnegie Mellon UniversityAssistant Professor
  2. NowGoogle DeepMindSenior Research Scientist
  3. GoogleResearch Scientist

Education

  1. 2022University of PennsylvaniaPhD

From the biography.

  1. 16% of AI career moves in the 2020s went from a university to a companyIncludes Daphne Ippolito. We estimate that 16% of AI career moves in the 2020s went from a university to a company; 29% of moves starting at a university went to a company, and 76% ended the university role.
  2. Google had the most people shaping AI every year from 2014 to 2025Includes Google. Our estimates of which organizations employed the most people shaping AI each year from 1955 to 2025, tracing the shift from university labs to companies.
  3. Google has the highest combined h-index among AI organizationsIncludes Google. We estimate that Google holds the highest combined h-index at 8,016, while universities account for 73% of the deduplicated total across AI organizations.
  4. 20% of people at companies who shape AI also hold a university role, led by Google and Stanford UniversityIncludes Google. We estimate that 539 of the 6,766 people with a known current organization (8%) hold both a university role and a company role, including 20% of people at companies and 13% of people at universities.

Colleagues

People named in this biography, and people with dated positions at the same organization in the same years.

Sources and credits

Photo: We have no published photo of Daphne Ippolito yet. Send us a free-licensed photo.

  1. Unsupervised Hierarchical Story Infilling aclanthology.org. Checked 22 October 2026.
  2. Automatic Detection of Generated Text is Easiest when Humans are Fooled aclanthology.org. Checked 22 October 2026.
  3. RoFT: A Tool for Evaluating Human Detection of Machine-Generated Text - ACL Anthology aclanthology.org. Checked 22 October 2026.
  4. Daphne Ippolito - MIT Rising Stars risingstars-eecs.mit.edu. Checked 22 October 2026.
  5. Daphne Ippolito - Language Technologies Institute - School of Computer Science - Carnegie Mellon University lti.cmu.edu. Checked 22 October 2026.
  6. Deduplicating Training Data Makes Language Models Better - ACL Anthology aclanthology.org. Checked 22 October 2026.
  7. [2211.05030] Creative Writing with an AI-Powered Writing Assistant: Perspectives from Professional Writers arxiv.org. Checked 22 October 2026.
  8. Generative AI Innovation Incubator cs.cmu.edu. Checked 22 October 2026.
  9. Scalable Extraction of Training Data from (Production) Language Models arxiv.org. Checked 22 October 2026.
  10. CyLab Seminar: Daphne Ippolito - CyLab Security and Privacy Institute cylab.cmu.edu. Checked 22 October 2026.
  11. Logistics - Large Language Model Applications / Spring 2026 cmu-llms.org. Checked 22 October 2026.