ANTHILLAsk your own question
Answer · confidence: partial

Who made the first rules limiting outside coders using a company's AI?

Research question: Who set the first published rules for how outside developers could use a company's AI system, and what limits did those rules impose?

The records do not show who published the first rules for outside developers, so this answer is partial. In 2020, OpenAI asked developer Jason Rohrer to bring his GPT-3 chatbot into line with its safety guidelines, but that was an enforcement request, not a published rule 1. The earliest dated published rules in the set are OpenAI's March 2023 usage policies, which allowed API integration into products on all major technology platforms and app stores, subject to use case restrictions 3. Those policies barred uses such as adult content, adult industries, and dating apps, and barred tailored financial advice without a qualified person's review and the diagnosis of health conditions or treatment instructions 57. Consumer-facing medical, financial, legal, and news uses had to disclose that AI was in use, and repeated or serious violations could lead to account suspension or termination 910. The newest fact in the set is dated March 23, 2026, from Google's Gemini terms, so the most recent evidence is about six months old as of October 2026 1214.

Checked 2026-10-19. Only documented and supported facts are used.
Dig deeper on this question

Sources

[1] supported
In 2020, OpenAI asked Jason Rohrer to change his GPT-3-based chatbot, which he had built on his GPT-3 account for the 'Project December' platform, so that it complied with OpenAI's safety guidelines.
"OpenAI asked for changes to the chatbot to comply with their safety guidelines; Rohrer disconnec"
Artificial general intelligence
[2] supported
In April 2023, OpenAI revised its terms of service so that businesses using its API to run their AI services could no longer include "GPT" in their product names or branding.
"In April 2023, OpenAI revised the brand guidelines in its terms of service to indicate that other businesses using its API to run their AI services would no longer be able to include "GPT" in such names or branding."
Generative pre-trained transformer
[3] documented
OpenAI's March 2023 usage policies allowed API integration into products on all major technology platforms and app stores, subject to use case restrictions.
"Subject to the use case restrictions mentioned below, we allow the integration of our API into products on all major technology platforms, app stores, and beyond."
What changed on openai.com/policies/usage-policies/ between 2023-03-10 and 2026-10-02
[4] supported
In May 2023, OpenAI hired a brand management service to notify its API customers of the naming policy, but the notices stopped short of overt legal claims such as trademark infringement allegations or cease-and-desist demands.
"In May 2023, OpenAI engaged a brand management service to notify its API customers of this policy, although these notifications stopped short of making overt legal claims"
Generative pre-trained transformer
[5] documented
The 2023 policy barred use of OpenAI's models for a list of activities, including adult content, adult industries, and dating apps.
"Adult content, adult industries, and dating apps"
What changed on openai.com/policies/usage-policies/ between 2023-03-10 and 2026-10-02
[6] supported
As of November 2023, OpenAI still barred its API licensees from naming their own products with "GPT".
"As of November 2023, OpenAI still prohibits its API licensees from naming their own products with "GPT""
Generative pre-trained transformer
[7] documented
The 2023 policy barred offering tailored financial advice without a qualified person's review, and diagnosing health conditions or providing treatment instructions.
"Offering tailored financial advice without a qualified person’s review Diagnosing a certain health condition, or providing treatment instructions"
What changed on openai.com/policies/usage-policies/ between 2023-03-10 and 2026-10-02
[8] supported
OpenAI allowed ChatGPT Plus subscribers to create custom GPTs, and its terms say subscribers may use "GPT" in their names, though this is discouraged.
"it has begun enabling its ChatGPT Plus subscribers to make "custom versions of ChatGPT" called GPTs on the OpenAI site."
Generative pre-trained transformer
[9] documented
Consumer-facing uses in medical, financial, and legal industries, and in news generation or summarization, had to show users a disclaimer that AI is being used and of its potential limitations.
"Consumer-facing uses of our models in medical, financial, and legal industries; in news generation or news summarization; and where else warranted, must provide a disclaimer to users informing them that AI is being used and of its potential limitations."
What changed on openai.com/policies/usage-policies/ between 2023-03-10 and 2026-10-02
[10] documented
Under the 2023 policy, repeated or serious violations could lead to suspension or termination of the offending account.
"Repeated or serious violations may result in further action, including suspending or terminating your account."
What changed on openai.com/policies/usage-policies/ between 2023-03-10 and 2026-10-02
[11] supported
Procore may set and enforce API usage limits, such as call counts or rate limits, at its sole discretion and without prior notice or liability, and may monitor usage.
"Procore may set and enforce limits on API access and use, for example numbers of calls, requests or rate limits, in its sole discretion and without prior notice or liability, and may monitor usage for quality, security and compliance."
Procore API Terms and AI Terms
[12] supported
Developers are barred from using the Gemini API or Google AI Studio to build competing models, and from reverse engineering, extracting, or copying any component of these services, including underlying data or model parameter weights.
"您不得使用这些服务来开发与这些服务(例如,Gemini API 或 Google AI Studio)竞争的模型,也不得尝试对这些服务中的任何组件 (包括底层数据或模型,例如参数权重)进行逆向工程、提取或复制。"
Gemini API Additional Terms of Service  |  Google AI for Developers
[13] supported
Developers may not use API data to train, retrain, fine-tune, or benchmark any machine learning or artificial intelligence system.
"may not use API Data to train, re-train, fine-tune or benchmark any machine learning or artificial intelligence system"
Procore API Terms and AI Terms
[14] documented
Developers may not use the services for clinical practice or medical advice, or in any way regulated by medical device authorities that requires their license or approval.
"您不得使用这些服务进行临床实践、提供医疗建议,或以任何受医疗设备监管机构监督或需要其许可或批准的方式使用这些服务。"
Gemini API Additional Terms of Service  |  Google AI for Developers
[15] supported
Users who violate or circumvent the stated prohibitions may lose access to OpenAI systems or face other penalties.
"Under these terms, users who violate or circumvent the stated prohibitions may lose access to OpenAI systems or face other penalties as described by the policy."
Usage Policies
[16] supported
Developers may not directly or indirectly charge customers for access to the Procore Services or APIs' functionality, though they may charge a standard overall fee for their applications.
"A developer may not “directly or indirectly charge Customers for use of, or access to, the functionality of the Procore Services or APIs”, although it may charge a standard overall fee for its Applications in accordance with the terms."
Procore API Terms and AI Terms