OpenAI Accuses Chinese Firm of Stealing AI Model Secrets
Left 50%
Center 0%
Right 50%
2 left · 0 center · 2 right
What happened
Cybersecurity firm Proofpoint released a report with evidence that Chinese hackers attempted to steal emails from artificial intelligence experts at American and Japanese universities and research organizations. Rep. Ro Khanna (D-Calif.), ranking member of the House Select Committee on the Chinese Communist Party, asked leading U.S. AI labs for information on Beijing’s alleged efforts to steal model weights from those firms. Separately, the winner of a cybersecurity contest prevailed by relying on an open-weight Chinese AI model.
Sourcing check · Mostly on the record · 4 articles
3 of 4 articles cite a named source for the story. It was on the record from the first report.
What these terms mean and how they're counted
- On the record · names a source 3 of 4
- On the record means the article cites a source by name for the story's main claim: a person, or an institution speaking officially, such as a company statement or a named agency's spokesperson. The opposite is an unnamed source. The count is how many of the counted articles cite at least one such source; the names are everyone named across them, each once however the articles spell it. Here: OpenAI, Thomas Duesterberg and Proofpoint
- Primary evidence · cites a document or data 2 of 4
- Primary evidence means the article cites a record you could check yourself: a court filing or ruling, a dataset or report, a recording or transcript, an official statement or press release, market data. A quote is not a document. The count is how many articles cite one. Here: company report, company statement, report
- Credits another outlet 0 of 4
- The article says the story comes from another outlet's reporting (“first reported by Reuters”). Those articles relay the story; they are not independent confirmation. Here: Each article presents its own reporting.
- Unnamed sources 0 of 4
- The main claim is attributed to people who are not identified: “people familiar with the matter”, “officials said”. The count is how many articles do this.
- How it developed
- When the first report landed and whether it named anyone; when the first named source and the first document appeared. Here: First report Oct 1, 02:14 UTC by CNBC, naming its source. First document 19 hours later, from Daily Caller (company report).
- Not counted 0
- Opinion, analysis and video pieces, and articles too short to assess. Every count above is out of the counted articles. An article republished under a new headline is one article.
This is about who is cited, not whether the story is true. Each source below says what it names. How we count
Omitted — what each side leaves out
Unpacked
The two sides describe different Chinese operations against AI, so a reader of either never learns the other's alleged target or even that a second operation was reported. Right-leaning coverage reports a Proofpoint finding that Chinese hackers tried to steal emails from AI experts at American and Japanese universities and research organizations. Left-leaning coverage instead reports Rep. Ro Khanna seeking information from leading U.S. labs on Beijing's alleged efforts to steal model weights.
Left-leaning coverage leads with a cybersecurity contest won by an open-weight Chinese model and a congressional information request rather than any named intrusion. Right-leaning coverage leads with a cybersecurity firm's evidence of email theft from university researchers.
Did any operation actually obtain model weights or source code?
Bottom line
Breitbart covers a Proofpoint report on email theft from university AI experts and never mentions Khanna. The Hill covers Khanna's request about stolen model weights and never mentions that email campaign.
The receipts — 2 facts one side's coverage lacks
- Khanna is the ranking member of the House Select Committee on the CCP
- A cybersecurity firm named Proofpoint issued a report providing evidence of the email-theft attempts
The Left View
Left-leaning coverage treats China’s AI challenge as both an open-model competition problem and a theft problem requiring oversight of U.S. labs. The New York Times presents the contest result as evidence of “the Power of China’s Approach to A.I.” and as “another sign that available A.I. systems could already be too powerful to easily regulate.” The Hill casts Khanna’s request as Democratic pressure on leading U.S. AI firms to disclose what they know about Beijing’s alleged efforts to obtain their code and weights.
The Right View
Breitbart frames the Proofpoint report as an accusation of Chinese cyber-espionage: hackers “attempting to steal emails” from AI experts, aimed at researchers and organizations in the United States and Japan rather than at open-model competition or regulation of already-released systems.
Our Take (balanced)
The left’s strongest argument is that China’s open-weight approach is already producing capable systems in real cybersecurity tasks, with the contest win cited as a sign that “available A.I. systems could already be too powerful to easily regulate.” The right’s strongest argument is that Chinese operators are actively targeting AI researchers’ private communications, backed by Proofpoint’s reported evidence of email-theft attempts. The central unresolved tension is whether the core problem is openly circulating Chinese models that outrun regulation, or clandestine efforts to steal Western experts’ emails and model weights.
3 sources
More stories
The week's bottom lines, in your inbox
One email a week: the five stories that mattered and what they actually mean. Free.