112:, they must not generate content that promotes overthrowing the socialist system, harms national security, or undermines social stability. They must not promote discrimination, terrorism, extremism, violence, obscenity, or false and harmful information prohibited by law. Models should be designed and trained to avoid discrimination or
136:
Article 6 promotes independent innovation in core AI technologies, infrastructure development, and data resource platforms. It encourages international cooperation, collaborative sharing of compute, expanding high-quality public training data sources, and the adoption of secure chips and resources.
163:
Article 11 obligates providers to protect user privacy, minimize data collection and retention of personal information, and avoid illegally giving user information to others. Providers should accept user requests to review, copy, modify, or delete their personal information.
173:
Article 14 obligates providers to promptly address when a service generates illegal content or if it is used by users to engage in illegal activities. This may include halting generation, retraining a model, warning users, and reporting to a relevant authority.
104:
Article 3 sets out key principles that will guide the regulatory approach: equal emphasis on development and security, promoting innovation along with rule of law, and adopting inclusive prudence with classified supervision of generative AI services.
93:
Article 1 outlines the intention of the measures: promoting the healthy development and regulated usage of generative AI, while safeguarding national security, societal public interests, and the legal rights of citizens and organizations.
148:
Article 8 requires providers to formulate clear, specific and operable data labeling rules in line with the measures. It mandates quality assessment, sampling verification, personnel training and supervision for labeling work.
140:
Article 7 mandates that providers employ effective measures to ensure that training data is high-quality, accurate, diverse, and legitimately sourced in a way that does not infringe upon intellectual property rights. If
132:
Article 5 encourages innovative application of generative AI across industries to create positive, healthy and high-quality content. It supports collaboration on AI innovation, data and application development.
157:
Article 9 legally establishes providers as producers of online content and processors of personal data. Providers should sign service agreements with users to define rights and obligations.
61:
160:
Article 10 mandates taking effective measures from excessive use by minors. Providers should clarify the appropriate use of their services to users and help users understand generative AI.
120:
rights and business ethics. They must respect others' rights and avoid endangering the physical and mental health of others. Effective measures should be taken to improve the
69:
327:
77:
65:
57:
185:
Article 16 gives regulatory oversight roles to cybersecurity, technology, industry, education, public security and other relevant government bodies.
200:
Article 21 defines the relevant authorities that may punish providers that violate the provisions. Crimes will be investigated according to law.
101:. It exempts internal services by companies, research institutions, and other organizations that do not provide services to the domestic public.
255:
121:
73:
53:
46:
279:
188:
Article 17 requires security assessments for services with public opinion influence, along with algorithm filing procedures.
98:
177:
Article 15 mandates complaint and reporting mechanisms for public grievances with timely processing and feedback.
108:
Article 4 requires generative AI services to adhere to relevant laws and respect social morality. Adhering to the
97:
Article 2 defines the scope of the measures as applying to public-facing generative AI services within the
117:
109:
231:
322:
142:
194:
Article 19 enables inspections of providers by officials, requiring cooperation and transparency.
113:
24:
211:
Article 23 requires obtaining relevant licenses to provide services where required by law.
306:
208:
Article 22 defines key terms such as generative AI, service providers, and service users.
256:"China finalizes first-of-its-kind rules governing generative A.I. services like ChatGPT"
197:
Article 20 allows the government to address foreign services that violate the measures.
167:
Article 12 requires labeling synthesized pictures, videos and other generated content.
316:
191:
Article 18 allows users to complain to authorities about non-compliant services.
170:
Article 13 requires stable and continuous services for normal usage by users.
214:
Article 24 establishes that the measures take effect on 15 August 2023.
232:"China just gave the world a blueprint for reigning in generative A.I."
32:
42:
49:
within the country. The measures took effect on 15 August 2023.
21:
Interim
Measures for the Management of Generative AI Services
38:
Shēngchéng shì réngōng zhìnéng fúwù guǎnlǐ zànxíng bànfǎ
181:
Section 4: Supervision, inspection, and legal liability
128:Section 2: Technology development and governance
56:, along with six other national regulators: the
284:Library of Congress, Washington, D.C. 20540 USA
70:Ministry of Industry and Information Technology
8:
78:National Radio and Television Administration
58:National Development and Reform Commission
280:"China: Generative AI Measures Finalized"
145:is used, the users must provide consent.
124:and accuracy of generative AI services.
223:
41:) are a set of measures introduced by
328:Regulation of artificial intelligence
16:Chinese regulation introduced in 2023
7:
204:Section 5: Supplementary provisions
66:Ministry of Science and Technology
54:Cyberspace Administration of China
47:generative artificial intelligence
14:
153:Section 3: Service specifications
52:The measures were issued by the
37:
1:
254:Kharpal, Arjun (2023-07-13).
89:Section 1: General provisions
74:Ministry of Public Security
344:
116:. Services should respect
99:People's Republic of China
45:to regulate public-facing
28:
118:intellectual property
110:Core Socialist Values
62:Ministry of Education
143:personal information
335:
294:
293:
291:
290:
276:
270:
269:
267:
266:
251:
245:
244:
242:
241:
228:
114:algorithmic bias
39:
30:
343:
342:
338:
337:
336:
334:
333:
332:
313:
312:
303:
298:
297:
288:
286:
278:
277:
273:
264:
262:
253:
252:
248:
239:
237:
230:
229:
225:
220:
206:
183:
155:
130:
91:
86:
29:生成式人工智能服务管理暂行办法
17:
12:
11:
5:
341:
339:
331:
330:
325:
315:
314:
311:
310:
302:
301:External links
299:
296:
295:
271:
246:
222:
221:
219:
216:
205:
202:
182:
179:
154:
151:
129:
126:
90:
87:
85:
82:
15:
13:
10:
9:
6:
4:
3:
2:
340:
329:
326:
324:
321:
320:
318:
308:
305:
304:
300:
285:
281:
275:
272:
261:
257:
250:
247:
236:
233:
227:
224:
217:
215:
212:
209:
203:
201:
198:
195:
192:
189:
186:
180:
178:
175:
171:
168:
165:
161:
158:
152:
150:
146:
144:
138:
134:
127:
125:
123:
119:
115:
111:
106:
102:
100:
95:
88:
83:
81:
79:
75:
71:
67:
63:
59:
55:
50:
48:
44:
40:
34:
26:
22:
309:(in Chinese)
287:. Retrieved
283:
274:
263:. Retrieved
259:
249:
238:. Retrieved
234:
226:
213:
210:
207:
199:
196:
193:
190:
187:
184:
176:
172:
169:
166:
162:
159:
156:
147:
139:
135:
131:
122:transparency
107:
103:
96:
92:
51:
36:
20:
18:
323:Chinese law
317:Categories
289:2023-07-24
265:2023-07-24
240:2023-07-24
218:References
307:Full text
84:Articles
235:Fortune
25:Chinese
76:, and
60:, the
35::
33:pinyin
27::
43:China
260:CNBC
19:The
319::
282:.
258:.
80:.
72:,
68:,
64:,
31:;
292:.
268:.
243:.
23:(
Text is available under the Creative Commons Attribution-ShareAlike License. Additional terms may apply.