Guide-yakabatana mibvunzo · Pakati Nhanho
China’s Cross-Border Data Rules and AI Training Quiz
Check China’s cross-border AI training rules for data classification, export thresholds, exemptions, and required transfer mechanisms.
Mubvunzo 1 ye 8
Zvipikirwa zve data zveChina 2024 zvakatanga rini kushanda?
Tarisa mubvunzo wese8 mibvunzo
Mhinduro uye tsananguro zvinoonekwa paunenge uchitora bvunzo.
- Zvipikirwa zve data zveChina 2024 zvakatanga rini kushanda?
- 22 March 2024
- 1 January 2023
- 1 August 2025
- 1 January 2026
- Nderipi gore rega rega rehuwandu husinganzwisisi ruzivo rwemunhu hunowanzo konzera kuongororwa kwekuchengetedza kune asiri weCIIO?
- At least 10,000 individuals
- Any transfer of one record
- At least one million individuals in a calendar year
- Exactly 100,000 individuals
- Ndeipi kondirakiti yakajairwa kana chitupa chinowanzodiwa paruzivo rwemunhu rusina hanya?
- Any amount below 100,000
- At least 100,000 but fewer than one million individuals, subject to exceptions
- At least one million
- Only when the data are important data
- Iwo mutemo unobata ese AI data yekudzidziswa seyakasunungurwa kubva kune ekunze zvinodzora?
- Yes, all model training is exempt
- Yes, if the data are used for research
- No, every dataset always needs security assessment
- No; teams must classify data and identify the applicable transfer route
- When may data not be treated as important data for assessment purposes?
- Whenever it is stored in a cloud
- Whenever it is used for training
- If authorities have not notified or publicly identified it as important data under the rules
- Whenever a company labels it synthetic
- Chii chingaite nzvimbo yemahara-yekutengeserana pasi pezvigadziro?
- Create an approved negative list for data-export mechanisms
- Waive all PIPL requirements
- Approve every AI model
- Exempt sensitive information everywhere
- What else may apply when personal information is exported?
- Mutemo wekodzero chete
- No other obligations below a threshold
- Only a cloud-security audit
- PIPL notices, separate consent, and impact-assessment duties as applicable
- Before training across borders, which step should a team take first?
- Count only total rows
- Map data subjects, sensitivity, volume, recipient, and transfer path
- Ask only where the model is trained
- Assume de-identification is complete