Kirin Chang

Research Fellow @USALINYU @ILI_NYU @NYULaw | Affiliate Research Fellow, AI and the Future of Work Program @AIandWork @EmoryLaw

[Kirin Chang @chengchi_chang]

If Sullivan & Cromwell (@sullcrom) can file a motion with 40+ hallucinated citations, the rest of us probably want a safety net. Just launched https://t.co/CafvGyTuLs: a free, open source tool (yes, really) to catch AI citation hallucinations. Demo attached.

[Kirin Chang @chengchi_chang]

@david__simon https://t.co/hIKQogpt00

[Lawrence Solum @lsolum]

Cheng-chi (Kirin) Chang (NYU), Algorithm as Agent: AI Vendor Liability for Title VII Hiring Discrimination After Loper Bright, on SSRN. The algorithm was always the agent. Highly recommended. https://t.co/bAGshsyQmK

[The Georgetown Law Journal @GeorgetownLJ]

Courts have reached for terms of negligence when sanctioning attorneys for submitting AI-generated citations to nonexistent cases. @chengchi_chang rejects this characterization and proposes a framework for the treatment of AI hallucinations. https://t.co/CUQ2ZoDCyh

[Stanford Law Review @StanLRev]

@chengchi_chang examines Mobley v. Workday, the most consequential AI discrimination case currently in litigation, and the impacts of Loper Bright on AI civil-rights litigation. https://t.co/ozrceB8mSJ

[Kirin Chang @chengchi_chang]

@akoustov This kinda already happened? AJPS (top-3 poli sci journal) dropped its ban on AI-drafted manuscripts on Aug 11: you can use AI for the whole writing process now, as long as you disclose and take responsibility (source: https://t.co/F5TKMH8G4m)

[Kirin Chang @chengchi_chang]

@danravicher People should use https://t.co/CafvGyTuLs, the completely free and open-source tool I built (yes, really!), to avoid this!

https://t.co/v0rbAXWAxN

[Kirin Chang @chengchi_chang]

If Sullivan & Cromwell (@sullcrom) can file a motion with 40+ hallucinated citations, the rest of us probably want a safety net. Just launched https://t.co/CafvGyTuLs: a free, open source tool (yes, really) to catch AI citation hallucinations. Demo attached.

[Kirin Chang @chengchi_chang]

@RobertFreundLaw People should use https://t.co/CafvGyTuLs, the completely free and open-source tool I built (yes, really!), to avoid this!

https://t.co/v0rbAXWAxN

[Kirin Chang @chengchi_chang]

If Sullivan & Cromwell (@sullcrom) can file a motion with 40+ hallucinated citations, the rest of us probably want a safety net. Just launched https://t.co/CafvGyTuLs: a free, open source tool (yes, really) to catch AI citation hallucinations. Demo attached.

[Kirin Chang @chengchi_chang]

@Doradu98 Thank you Zheng!

[Kirin Chang @chengchi_chang]

If Sullivan & Cromwell (@sullcrom) can file a motion with 40+ hallucinated citations, the rest of us probably want a safety net. Just launched https://t.co/CafvGyTuLs: a free, open source tool (yes, really) to catch AI citation hallucinations. Demo attached.

[Kirin Chang @chengchi_chang]

@bnafOg @sullcrom Thanks, totally agree. I built https://t.co/CafvGyTuLs as a verification step, not a generator. It checks each citation against real legal databases before filing

[Kirin Chang @chengchi_chang]

@bnafOg @sullcrom Great point, and yes, I’ve seen that exact failure mode. I built https://t.co/CafvGyTuLs to cross-check case names and citation metadata (volume, reporter, page, year), so plausible but wrong citations get flagged instead of passing through.

[Kirin Chang @chengchi_chang]

Huge thanks to @FreeLawProject for the open legal infrastructure behind this. No idea if this reaches the people who need it most, but if it reaches you and there's a feature you want, drop it below. More replies = more motivation to keep shipping after work!

[Kirin Chang @chengchi_chang]

Full disclosure: this is also the practical side of a forthcoming piece. Hallucinated Authority: AI Citations as Reckless Misrepresentation, 115 @GeorgetownLJ Online (forthcoming 2026). It argues unverified AI cites can be reckless misrepresentation, not routine error.

[Kirin Chang @chengchi_chang]

Thrilled to share "AI, Legal Labor, and the Jevons Paradox," forthcoming in @nyulawreview.

Jevons paradox: efficiency gains expand consumption. In law: cheaper AI surfaces more demand, more demand means more work. The profession expands, not shrinks.

https://t.co/rx2V77ROVc

[Kirin Chang @chengchi_chang]

This is a wonderful opportunity to spend a year conducting research @nyulaw while engaging with an international community of legal scholars and practitioners. If you have any questions about the program or application process, please feel free to reach out!

[U.S.-Asia Law Institute @USALINYU]

The U.S.-Asia Law Institute at @nyulaw is accepting applications for 2026-2027 visiting scholars. Open to legal scholars and practitioners with LLM/JD or equivalent.

Deadline: Jan 31, 2026

More info: https://t.co/U6Y9ym6LHe

ご応募お待ちしています！

[Kirin Chang @chengchi_chang]

New draft book chapter: "Superintelligent AI and the Right to Be Forgotten" for Research Handbook on Law in an Age of Superintelligence (@ElgarPublishing , forthcoming).

Grateful to editors Woody Barfield & Marc Blitz for the invitation.

Draft: https://t.co/jbRz6MgvYh

[Kirin Chang @chengchi_chang]

My article "Wires of War" is published in @UPennJIL!

Submarine cables carry 95% of global data but face hybrid warfare threats. Ukraine & Taiwan show vulnerabilities to sabotage & cyberattacks.

Solutions: treaties, cooperation & partnerships.

Read more: https://t.co/Yq9UjMK0XN

[Lawrence Solum @lsolum]

Chang on the Framework Convention on Artificial Intelligence and Human Rights, https://t.co/OlRPuPt6eW - Cheng-chi (Kirin) Chang (Emory University School of Law) has posted The First Global AI Treaty: Analyzing the Framework Convention on Artificial Intelligence and the Eu AI Act (2024 University of Illinois Law Review (Online) 86 (2024)) on SSRN.