Policy & Regulation

Judge Sets January Trial in Former OpenAI Engineer Bugging Case

U.S. District Judge Vince Chhabria set a January 20, 2027 trial for former OpenAI engineer Abdulrahman Alfozan and set deadlines for FBI reports, missing evidence and any Claude review.

Judge Sets Expedited Trial for Former OpenAI Engineer in Bugging Case
Judge Sets Expedited Trial for Former OpenAI Engineer in Bugging CasePexels / pixabay
By Rebecca Stone6 min read

Why it matters

  • Trial set for January 20, 2027 before U.S. District Judge Vince Chhabria in San Francisco.
  • Prosecutors must provide all FBI reports by November 24, 2026 and move promptly on other discovery.
  • Government must justify its objection to Claude review by November 20, 2026.
  • Nick Parker said investigators collected two laptops, a tablet, another phone and 160 hours of recordings.
  • The October 27, 2026 indictment carries a maximum five-year prison term, $250,000 fine and three years of supervised release.

By Rebecca Stone

CNBC reported on November 18, 2026, that U.S. District Judge Vince Chhabria set a January 20, 2027 trial for Abdulrahman Alfozan at a November 17, 2026 hearing in San Francisco. The former OpenAI engineering lead faces one federal charge over an alleged recording-device plant in an executive conference room.

Chhabria ordered prosecutors to disclose all FBI reports by November 24, 2026, and directed the government to move promptly on other trial discovery. Within seven days, the judge also asked to see any evidence that Alfozan had engaged in economic espionage, trade-secret theft or acting as an unregistered foreign agent.

Defense attorney David Rizk had sought the speedy trial after arguing that prosecutors had been slow to turn over evidence. Assistant U.S. Attorney Nick Parker said he was unaware of any evidence in those three areas. “At this point, I’m not suggesting the defendant stole anything,” Parker said.

The stakes extend beyond the bugging charge because Rizk’s proposal could place OpenAI’s potentially proprietary data inside a competing system under a court-imposed protective order. The court must balance prompt defense access against strict control of that material, according to Chhabria’s actions and CNBC’s report.

How did Claude enter the evidence dispute?

Rizk intends to use Anthropic’s Claude to review evidence. Parker told the court that the plan raised a serious concern because one seized laptop might contain proprietary OpenAI data, according to CNBC’s account of the hearing.

Chhabria expressed less concern and ordered prosecutors to file a motion by November 20, 2026, explaining why using a competing AI tool to analyze evidence might create a problem. His deadline makes the government’s rationale the next point of legal review.

Chhabria granted the stipulated protective order on November 3, 2026. It bars anyone authorized to see protected information or confidential material from exposing that material to an AI tool without prior written government consent, according to the order.

The measure covers large language models, generative AI services and AI-assisted software, whether cloud-based or otherwise. The order does not name Claude, even though Rizk has identified Anthropic’s model as the proposed tool.

For a proposed tool such as Claude, the order functions like a third-party records-custody rule: consent and certifications govern receipt, retention, training, third-party access and deletion.

To obtain consent, defense counsel must identify the tool and certify that it does not and will not retain or use the material for model training or expose it to unauthorized third parties. Counsel must also certify reasonable confidentiality safeguards and deletion of all such material when the case ends, according to the November 3 order.

The order separately prohibits publicly accessible AI systems that retain submissions for training. That restriction narrows the available consent path for any defense review that would send protected discovery to a public service, according to Chhabria’s order.

Chhabria also designated all source code on storage media and certain charts as confidential trade-secret material. Alfozan retains the right to challenge those designations, according to the November 3 order.

In a November 16, 2026 status statement, Rizk asked the court to strike the AI clause until the government justified it. Rizk argued that prosecutors should not police defense discovery practices, according to his filing.

He accepted the order to accelerate disclosure while reserving his objection. Rizk also voluntarily disclosed some AI use to prosecutors, according to the same status statement.

The protective order therefore does not settle whether Claude can examine the evidence. It instead makes consent, retention, training, third-party access, confidentiality and deletion the conditions for any authorized review, according to the order and Rizk’s request.

What evidence has the defense received?

Rizk described the government’s first production on November 9, 2026, in the November 16 filing. It contained:

  • Residence photographs.
  • OpenAI investigative and employment records.
  • Roughly 20 pages of FBI reports.
  • Limited surveillance footage.
  • Subpoena returns.

Rizk said the production did not include forensic device images, the alleged recordings or formal forensic and technical reports. Those materials were among the specified evidence his side requested by 9 a.m. on November 19, 2026, ahead of a renewed bail request.

Chhabria’s November 24 deadline for all FBI reports falls after that defense request. The two orders set distinct checkpoints: a defense deadline for additional evidence and a judicial deadline for the government’s report production, according to the hearing and Rizk’s filing.

At the November 17 hearing, Parker said investigators remained locked out of Alfozan’s phone. Investigators had collected another phone, a tablet and two laptops, along with 160 hours of conference-room recordings, according to Parker.

Parker described the recording device as encrypted. He also said data backed up to another device could have been accessible, according to CNBC’s report.

The government’s possession of the recordings does not place them in the defense’s hands. Rizk’s filing says the first production omitted the alleged recordings, while Parker said prosecutors held them and were still seeking access to material on Alfozan’s phone.

What does the indictment allege?

An October 27, 2026 indictment charges Alfozan with one count of interception of oral communications. The Justice Department alleges conduct at OpenAI in San Francisco from no later than September 28, 2026 through October 5, 2026.

Alfozan pleaded not guilty on November 6, 2026, and the court remanded him to custody, according to CNBC’s report of the case. The prosecution’s allegations have not been proven, and Alfozan is presumed innocent.

In the Justice Department’s original account, OpenAI personnel found an encrypted recorder concealed in toilet paper on the back of an executive conference-room television on September 30, 2026. The department said personnel found a second device in Alfozan’s bag on October 5.

The charge carries a maximum of five years in prison, a $250,000 fine and three years of supervised release, according to the October 27 indictment.

What deadlines come next?

Chhabria has set four dates for the next stage:

  • November 19, 2026, 9 a.m.: Defense request for forensic images, alleged recordings, formal reports and other specified evidence, ahead of a renewed bail request.
  • November 20, 2026: Government motion explaining why Claude review of evidence presents a problem.
  • November 24, 2026: Deadline for prosecutors to provide all FBI reports.
  • January 20, 2027: Trial date.

The immediate test comes on November 20, when prosecutors must explain why the defense’s proposed use of Claude would create a legal problem. By January 20, 2027, the court must reconcile defense access to 160 hours of recordings and missing forensic material with safeguards for proprietary OpenAI data.

The November 20 filing will show how far those controls extend. The bugging allegation remains unproven, and Alfozan remains presumed innocent.

Source: unite.ai

Share this article:

More from Rebecca Stone

Rebecca Stone

Show full bio

Correspondent covering companies, deals and strategy in artificial intelligence and machine learning at AI In Context.

728 articles

Related articles

  1. FBI Charges OpenAI Engineer With Hiding Recorder in Executive Room
  2. FBI says accused OpenAI spy kept folder labeled "Saudi Gov Docs"
  3. Feds Say OpenAI Engineer Hid a $13.99 Bug in an Executive Boardroom
  4. FBI arrests Saudi OpenAI employee over hidden conference room recorder
  5. Judge Orders Ex-OpenAI Engineer Accused of Bugging Exec Room to Stay Jailed

« Previous articleNext article »