Draft:Heidy Khlaaf
 | Review waiting, please be patient.
This may take 5 weeks or more, since drafts are reviewed in no specific order. There are 2,571 pending submissions waiting for review.
Where to get help
How to improve a draft
You can also browse Wikipedia:Featured articles and Wikipedia:Good articles to find examples of Wikipedia's best writing on topics similar to your proposed article. Improving your odds of a speedy review To improve your odds of a faster review, tag your draft with relevant WikiProject tags using the button below. This will let reviewers know a new draft has been submitted in their area of interest. For instance, if you wrote about a female astronomer, you would want to add the Biography, Astronomy, and Women scientists tags. Editor resources
Reviewer tools
|
Heidy Khlaaf | |
|---|---|
| Education | Florida State University (BS) University College London (PhD) |
| Known for | AI safety and assurance, military AI safety |
| Awards | TIME100 AI (2025) MIT Technology Review Innovators Under 35 (2025) |
| Scientific career | |
| Fields | artificial intelligence, safety engineering, computer security, formal verification |
| Workplaces | AI Now Institute Trail of Bits OpenAI |
| Thesis | The Past, Present, and Future(s): Verifying Temporal Software Properties |
| Website | www |
Heidy A. Khlaaf is a computer scientist and safety engineer whose work focuses on the safety, security and assurance of artificial intelligence systems.[1] She is the chief AI scientist at the AI Now Institute, where her research concerns the use of AI in defense and national-security applications, including autonomous weapons and military decision-support systems.[2][3]
Khlaaf previously worked at OpenAI, where she contributed to the safety assessment of Codex, and at cybersecurity firm Trail of Bits, where she led work on AI and machine-learning assurance.[4][5]
In 2025, Time included Khlaaf in its TIME100 AI list, and MIT Technology Review selected her for its global Innovators Under 35 list for her research on the safety and security of AI systems used in military applications.[2][3]
Education and early career
[edit]Khlaaf earned a Bachelor of Science in computer science and philosophy from Florida State University and a doctorate in computer science from University College London.[1] Her doctoral research concerned formal verification, temporal logic and automated methods for verifying properties of infinite-state software systems.[6]
Artificial intelligence safety and assurance
[edit]Khlaaf worked as a senior systems safety engineer at OpenAI in 2021.[4] During the development of Codex, she contributed to work examining the risks associated with large language models capable of generating computer code.
Khlaaf, Pamela Mishkin, Joshua Achiam, Gretchen Krueger and Miles Brundage subsequently published A Hazard Analysis Framework for Code Synthesis Large Language Models, which proposed applying structured hazard-analysis methods to code-generating language models.[7] Their risk assessment methodology is now widely used across the AI industry.[2]
Military and national-security AI
[edit]As chief AI scientist at the AI Now Institute, Khlaaf has focused on the safety and security implications of deploying generative and foundation models in military and national-security systems.[3][2]
In 2024, Khlaaf, Sarah Myers West and Meredith Whittaker published Mind the Gap: Foundation Models and the Covert Proliferation of Military Intelligence, Surveillance, and Targeting. The paper argued that policy discussions about national-security risks from foundation models had focused heavily on their potential contribution to chemical, biological, radiological and nuclear weapons while devoting less attention to their use in military intelligence, surveillance and targeting systems.[8]
Khlaaf's work has emphasized the reliability standards and risk thresholds applied to AI systems used in military settings. MIT Technology Review, in selecting her as an Innovator Under 35, highlighted her comparisons between the assurance standards traditionally imposed on defense systems and the lower measured reliability of some experimental AI targeting technologies.[3]
In 2025, Khlaaf and Myers West published Safety Co-Option and Compromised National Security: The Self-Fulfilling Prophecy of Weakened AI Risk Thresholds. They argued against weakening established safety and security requirements in order to accelerate military adoption of artificial intelligence and called for decisions about acceptable risk to be subject to public and regulatory oversight.[9]
Time cited this work in its 2025 TIME100 AI profile of Khlaaf and described her as a critic of arguments that an international AI arms race justified rapid deployment of generative AI in military systems.[2] Her views on the integration of generative AI into military technology have also been featured by The Verge.[10]
See also
[edit]References
[edit]- 1 2 Wiggers, Kyle (10 March 2024). "Women in AI: Heidy Khlaaf, safety engineering director at Trail of Bits". TechCrunch. Retrieved 8 August 2026.
- 1 2 3 4 5 Booth, Harry (26 August 2025). "Heidy Khlaaf: The 100 Most Influential People in AI 2025". Time. Retrieved 8 August 2026.
- 1 2 3 4 "Heidy Khlaaf". Innovators Under 35. MIT Technology Review. 2025. Retrieved 8 August 2026.
- 1 2 Smith, Tim (20 February 2023). "Europe's OpenAI alumni". Sifted. Retrieved 8 August 2026.
- ↑ Newman, Lily Hay; Burgess, Matt (16 January 2024). "A Flaw in Millions of Apple, AMD, and Qualcomm GPUs Could Expose AI Data". Wired. Retrieved 8 August 2026.
- ↑ "The Past, Present, and Future(s): Verifying Temporal Software Properties". UCL Discovery. University College London. 2018. Retrieved 8 August 2026.
- ↑ Khlaaf, Heidy; Mishkin, Pamela; Achiam, Joshua; Krueger, Gretchen; Brundage, Miles (2022). "A Hazard Analysis Framework for Code Synthesis Large Language Models". arXiv:2207.14157 [cs.AI].
- ↑ Khlaaf, Heidy; Myers West, Sarah; Whittaker, Meredith (2024). "Mind the Gap: Foundation Models and the Covert Proliferation of Military Intelligence, Surveillance, and Targeting". arXiv:2410.14831 [cs.CY].
- ↑ Khlaaf, Heidy; Myers West, Sarah (2025). "Safety Co-Option and Compromised National Security: The Self-Fulfilling Prophecy of Weakened AI Risk Thresholds". arXiv:2504.15088 [cs.CY].
- ↑ Field, Hayden (25 September 2025). "How AI safety took a backseat to military money". The Verge. Retrieved 8 August 2026.
External links
[edit]- Official website
- Heidy Khlaaf at the AI Now Institute
Category:Living people Category:Artificial intelligence researchers Category:Formal methods people Category:Women computer scientists Category:University College London alumni Category:Florida State University alumni Category:OpenAI people