Realtime AI News
Zhipu rolls out its strictest data protection mechanism after data-transfer controversy
Zhipu has brought online a data protection mechanism that Lianhe Zaobao describes as the strictest in China, following a controversy over alleged covert data transfers. The move puts model providers' data handling under scrutiny, and how verifiable the mechanism turns out to be will decide whether it restores trust.
Zhipu has launched a data protection mechanism after becoming embroiled in a controversy over alleged covert data transfers, according to a report by Lianhe Zaobao, which describes the mechanism as the strictest in China. The significance lies less in its technical design than in a leading model developer publicly adjusting its data practices once questions were raised.
The controversy centres on allegations that data was transmitted or used without users' knowledge. For companies selling cloud-based model services, that is the most sensitive ground there is: customers want to know how prompts are stored, whether they feed training runs, and who can access them.
Calling the mechanism the strictest in China is a strong self-positioning. Chinese regulation already imposes clear requirements on data security and personal information protection, so stacking further commitments on top both answers public criticism and supplies material for enterprise sales, where procurement teams routinely interrogate data flows and retention.
Statements and verification are different things. Public information so far is limited to the report's characterisation; it does not say which parts of the pipeline the mechanism covers, who audits it, or whether it is open to third-party checks. For enterprise buyers, independent verification usually matters more than the word strict.
The episode also reflects a common position for Chinese model vendors at the commercialisation stage. As capability gaps narrow, data governance becomes a hard gate in purchasing decisions, and any vendor caught in a data dispute can see rivals turn the issue into a sales advantage.
What to watch next is whether Zhipu publishes the mechanism's substance and its audit arrangements, whether regulators say anything on the record, and whether competitors follow with comparable commitments. If mechanisms like this turn into a comparable industry standard, a short-lived controversy becomes a long-term rule of competition.
Why it matters
Data governance is becoming a hard gate in model procurement, so Zhipu's move doubles as crisis response and trust rebuilding for enterprise sales. With no published detail or third-party audit, its real effect remains to be seen.
Nearby Updates
All09/21, 09:07
Plugin4Shell zero-click RCE hits four AI coding agents
A report captured through Google News describes a zero-click remote code execution vulnerability called Plugin4Shell that affects four AI coding agents. Zero-click means the flaw needs no extra action from a user in the interface once its trigger condition is met.
09/21, 07:38
StepFun's Step 5 Preview claims a global open-source top-three spot at one-eighth the cost of Claude Opus 5
According to a report on mrjjxw.com, StepFun has released Step 5 Preview, claiming it has entered the global open-source top three and that its cost per task is just one-eighth that of Claude Opus 5. The same report notes that beyond leaderboard placement, real-world deployment still requires work.
09/21, 02:41
The Next Web: Only Z.ai Can Open the Code Z.ai Uploaded
The Next Web reports that code uploaded by Z.ai can only be opened by Z.ai itself, shifting the focus from the act of publishing to what anyone else can actually use afterwards. The account raises the question of whether a release is readable and verifiable by anyone other than the company that posted it.
09/21, 01:17
Huawei lays out a supernode and optical interconnect compute blueprint
Shanghai Securities News reports that Huawei has presented a new computing blueprint built around supernodes and optical interconnect. The pairing targets the bandwidth and power bottlenecks that increasingly decide how far AI clusters can scale.