Answer · confidence: partial
What rules limit who can use AI, like age limits or bans, and where?
Research question: What current government rules, age limits, or bans now restrict who can use AI tools, and where do they apply?
China's anthropomorphic AI rules are the only enacted restriction in the verified facts that targets minors, and they apply to emotionally interactive AI services within mainland China from July 15, 2026 2. Providers must implement time-usage controls and avoid content that induces minors toward unhealthy habits 85. China's 2023 generative AI measures already required providers to take steps to stop minor users from over-relying on or becoming addicted to these services 3. In the United States, the GUARD Act, which would ban minors from AI companion tools, is only a proposal and not an enacted rule 1. The newest verified fact is the July 15, 2026 effective date, about three months before October 2026, so the facts cannot confirm whether later changes have occurred 2.
Checked 2026-10-14. Only documented and supported facts are used.
Dig deeper on this questionSources
[1] supported
The GUARD Act, proposed in December 2025, would prohibit minors from using AI companion tools. It is a proposal, not an enacted rule.
"The GUARD Act (Guidelines for User Age-verification and Responsible Dialogue) proposed in December 2025 aims to prohibit minors from using"2026 State and Federal AI Legislation Updates - Center for Democracy and Technology
[2] supported
The measures apply to anthropomorphic AI services that simulate a natural person's personality and provide sustained emotional interaction within mainland China.
"明确立法依据及适用范围（为中国大陆模拟自然人人格的持续性情感互动服务）"人工智能拟人化互动服务管理暂行办法
[3] supported
The measures require AI service providers to clarify the intended users, occasions and purposes of their service, and to take effective steps to prevent minor users from over-relying on or becoming addicted to generative AI services.
"第十条规定，服务提供者应明确服务的适用对象、场合和用途，并采取有效措施防范未成年人用户过度依赖或者沉迷生成式人工智能服务。"生成式人工智能服务管理暂行办法
[4] supported
Italy's data protection regulator (the Garante) imposed restrictions on ChatGPT, and the paper argues its approach overlooked the distinctive nature of generative AI.
"The Italian regulator adopted an approach that overlooked the distinctive nature of generative AI when imposing restrictions on ChatGPT."Theorizing the regulation of generative AI: lessons learned from Italy's ban on ChatGPT (2024)
[5] supported
The rules set service red lines that prohibit generating content that endangers national security, harms physical and mental health, leaks secrets, or induces minors toward unhealthy habits.
"划定服务红线，禁止生成危害国家安全、损害身心健康、泄露秘密及诱导未成年人不良嗜好等内容"人工智能拟人化互动服务管理暂行办法
[6] documented
In October 2022 the Commerce Department's BIS imposed controls on certain destinations, restricting access by foreign military and intelligence services to advanced computing chips and certain chip-manufacturing equipment.
"In October 2022, BIS imposed a first set of controls to certain destinations with the aim of preventing certain foreign military and intelligence services from obtaining the ability to indigenously develop advanced weapons modeling and advanced AI capabilities, by imposing restrictions on their access to advanced computing ICs and certain semiconductor manufacturing equipment (SME)"Framework for Artificial Intelligence Diffusion (Commerce Department, Industry and Security Bureau, 2025-01-15)
[7] supported
The measures apply to generative AI services that offer text, images, audio, video or other content to the public within mainland China.
"本办法适用于利用生成式人工智能技术向中华人民共和国境内公众提供文本、图片、音频、视频或其他内容的服务（第二条）。"生成式人工智能服务管理暂行办法
[8] supported
Service providers must implement time-usage controls, labeling of AI-generated content, and data security and personal information protection duties.
"要服务提供者落实人工智能生成合成内容标识义务、时长管控、数据安全与个人信息保护责任"人工智能拟人化互动服务管理暂行办法
[9] documented
The FCC proposed an exemption from TCPA restrictions for calls by people with speech or hearing disabilities using any technology, including AI, to communicate over the phone, as long as the calls contain no telemarketing or advertising content. This is a proposed rule, not a current one.
"Calls made by individuals with speech or hearing disabilities using any technology, including artificial intelligence (AI) technologies, designed to facilitate the ability of such individuals to communicate using an artificial or prerecorded voice over the telephone, provided that the calls must not include any telemarketing or advertising content."Implications of Artificial Intelligence Technologies on Protecting Consumers From Unwanted Robocalls and Robotexts (Federal Communications Commission, 2024-09-10)
[10] supported
The measures do not apply to organizations that only develop or use generative AI internally without offering services to the domestic public, such as industry associations, enterprises and research institutions.
"不适用于未向境内公众提供服务、仅在内部开发或使用生成式人工智能的组织，如行业协会、企业和研究机构。"生成式人工智能服务管理暂行办法
[11] supported
General-purpose customer service, knowledge Q&A, and work assistant AI tools are excluded because they do not involve sustained emotional interaction.
"一般性的智能客服、知识问答、工作助手类服务不涉及持续性情感互动，不适用本办法。"人工智能拟人化互动服务管理暂行办法
[12] supported
Validated End-User (VEU) status for entities depends on where they are headquartered: entities in Macau or Country Group D:5 destinations are excluded from eligibility under the standard path, while other destinations are eligible if they meet the applicable standards.
"Entities in destinations except Macau, destinations specified in Country Group D:5, or those listed in paragraph (a) to Supplement No. 5 to Part 740 are eligible if they meet the appropriate standards outlined in Supplement No. 10 to part 748"Framework for Artificial Intelligence Diffusion (Commerce Department, Industry and Security Bureau, 2025-01-15)
[13] documented
The FCC proposed requiring callers making AI-generated calls to give clear and conspicuous disclosure, when obtaining prior express consent, that they intend to use AI-generated voice or text content. This is a proposed rule, not a current one.
"Callers making an AI-generated call subject to the requirements contained in paragraphs (a)(1) through (3) of this section must provide clear and conspicuous disclosure that they intend to use AI-generated voice or text content on such calls when obtaining the prior express consent of the called party."Implications of Artificial Intelligence Technologies on Protecting Consumers From Unwanted Robocalls and Robotexts (Federal Communications Commission, 2024-09-10)
[14] supported
The final version narrowed the scope from the draft by limiting regulated activity to providing generative AI services to the public within mainland China.
"将监管对象限定为“向中华人民共和国境内公众提供生成式人工智能服务”的行为，并删除了部分针对模型训练与数据来源的绝对责任条款。"生成式人工智能服务管理暂行办法
[15] supported
Violations can be punished with warnings, orders to correct, suspension of service, or fines, with severe cases leading to service termination and heavier penalties.
"明确违规行为的处罚措施，包括警告、责令改正、暂停服务、罚款等，情节严重可停止服务并从重处罚。"人工智能拟人化互动服务管理暂行办法
[16] documented
Under current FCC rules, callers are not required to disclose before a call what technology is used in artificial or prerecorded voice calls.
"The Commission notes that its rules do not require pre-call disclosures about the technology used in artificial or prerecorded voice calls."Implications of Artificial Intelligence Technologies on Protecting Consumers From Unwanted Robocalls and Robotexts (Federal Communications Commission, 2024-09-10)
[17] supported
The current requirement for artificial or prerecorded voice calls is only that callers disclose their identity, not whether the call is AI-generated.
"which is simply that callers must disclose their identity when making an artificial voice or prerecorded call, and not specifically whether the call is an AI-generated call"Implications of Artificial Intelligence Technologies on Protecting Consumers From Unwanted Robocalls and Robotexts (Federal Communications Commission, 2024-09-10)
[18] supported
The TCPA does not define the terms 'artificial' or 'prerecorded voice,' so the FCC asked whether it could define them to exclude assistive technologies for people with disabilities.
"The Commission notes that the TCPA does not define the terms ``artificial'' or ``prerecorded voice.''"Implications of Artificial Intelligence Technologies on Protecting Consumers From Unwanted Robocalls and Robotexts (Federal Communications Commission, 2024-09-10)
[19] supported
Zhipu's chatbot Zhipu Qingyan (ChatGLM) passed the filing (备案) required under China's Interim Measures for the Administration of Generative AI Services on August 31, 2023, before it launched.
"2023年8月31日，智谱推出的 人工智能 大语言模型 “ 智谱清言 （ChatGLM）”通过了 中华人民共和国 《 生成式人工智能 服务管理暂行办法》备案，备案号为“Beijing-ChatGLM-20230821”"智谱
[20] supported
In January 2025 the US Commerce Department's Bureau of Industry and Security added Zhipu to its Entity List, which subjects the company to trade restrictions because the US treats listed companies as national security concerns.
"并宣布将智谱列入 實體清單 [ 13 ] ，该清单上的公司被美国视为涉及 國家安全 问题，而受到贸易限制 [ 1 ] [ 14 ]"智谱
[21] supported
The SEC's proposed conflicts rules would restrict firms' use of complex covered technologies, which lack explainability, to cases where all of the rule's requirements are met.
"The proposed conflicts rules would apply equally to these covered technologies, and firms would only be able to continue using them where all requirements of the proposed conflicts rules are met, including the requirements of paragraph (c)."Conflicts of Interest Associated With the Use of Predictive Data Analytics by Broker-Dealers and Investment Advisers (Securities and Exchange Commission, 2023-08-09)
[22] supported
Firms using covered technology to give advice or make recommendations are already subject to existing standards of conduct, and the proposed conflicts rules would apply on top of those standards.
"The proposed conflicts rules would apply in addition to these standards of conduct and other regulatory obligations."Conflicts of Interest Associated With the Use of Predictive Data Analytics by Broker-Dealers and Investment Advisers (Securities and Exchange Commission, 2023-08-09)
[23] supported
The CREATe researchers list age restrictions as one of the key things users need to know about generative AI products, alongside content restrictions, enforcement procedures, dispute resolution and copyright rules.
"Our research identifies the following key knowledge requirements: age and content restrictions, enforcement procedures, dispute resolution mechanisms, and copyright regulation for the input"Priorities for Generative AI Regulation in the UK: CREATe response to the Digital Regulation Cooperation Forum (DRCF) (2023)