OpenAI: Difference between revisions

(3 intermediate revisions by one other user not shown)

Line 26:

===Web Crawlers ignoring robots.txt (2025)===

In 2025, Jonathan Bailey from PlagiarismToday posted an article going into how ChatGPTs web crawlers were ignoring the sites Robots.txt file.<ref>https://www.plagiarismtoday.com/2025/07/23/chatgpt-ignores-robots-txt-rehashes-my-column/ ([http://web.archive.org/web/20260106080839/https://www.plagiarismtoday.com/2025/07/23/chatgpt-ignores-robots-txt-rehashes-my-column/ Archived])</ref> PlaigarismToday had blocked OpenAI's web crawlers in August of 2023, yet the latest ChatGPT model at the time provided data from articles that were posted the day before on the website, even though OpenAI wasn't supposed to be scraping ~~these web pages~~. This can be problematic for smaller websites, due to OpenAI's aggressive approach to web crawling, with their crawlers reportedly in a single week sending in more than 29 thousand requests to a wiki known as The Cutting Room Floor.

In 2025, Jonathan Bailey from PlagiarismToday posted an article going into how ChatGPTs web crawlers were ignoring the sites Robots.txt file.<ref>https://www.plagiarismtoday.com/2025/07/23/chatgpt-ignores-robots-txt-rehashes-my-column/ ([http://web.archive.org/web/20260106080839/https://www.plagiarismtoday.com/2025/07/23/chatgpt-ignores-robots-txt-rehashes-my-column/ Archived])</ref> PlaigarismToday had blocked OpenAI's web crawlers in August of 2023, yet the latest ChatGPT model at the time provided data from articles that were posted the day before on the website, even though OpenAI wasn't supposed to be scraping those webpages. This can be problematic for smaller websites, due to OpenAI's aggressive approach to web crawling, with their crawlers reportedly in a single week sending in more than 29 thousand requests to a wiki known as The Cutting Room Floor.

===ChatGPT Atlas and prompt-injection vulnerability (2025)===

In 2025, Brave posted an article about vulnerabilities that have agentic web browsers, such as ChatGPT Atlas, that consists of adding hidden malicious prompts in files, text or another media. Those prompts, combined with weak safeguards of the AI agents, can make them to expose and leak sensitive data of the user.<ref>https://owasp.org/www-community/attacks/PromptInjection ([http://web.archive.org/web/20260210124436/https://owasp.org/www-community/attacks/PromptInjection Archived])</ref>

===Funding of the Parents & Kids Safe AI Act and creation of a child safety organization (2026)===

In January 2026, OpenAI partnered with Common Sense Media and funded approximately $10 million dollars to support the California [[Parents & Kids Safe AI Act]] bill. The bill proposes to add [[age verification]], improvement of parental controls, and to prohibit [[Personalized ads|targeted advertising]] to underage users. On January 8, OpenAI created an organization, named Parents & Kids Safe AI Coalition.

In March 2026, the OpenAI's organization sent emails to child safety coalitions, only mentioning it was sponsored by Common Sense Media. The purpose of these e-mails was to convince the child safety coalitions to support the bill.<ref>{{Cite web |last=Dellinger |first=AG |date=1 Apr 2026 |title=Group Pushing Age Verification Requirements for AI Turns Out to Be Sneakily Backed by OpenAI |url=https://gizmodo.com/group-pushing-age-verification-requirements-for-ai-turns-out-to-be-sneakily-backed-by-openai-2000741069 |url-status=live |archive-url=https://web.archive.org/web/20260401225238/https://gizmodo.com/group-pushing-age-verification-requirements-for-ai-turns-out-to-be-sneakily-backed-by-openai-2000741069 |archive-date=2026-04-01 |access-date=4 Apr 2026 |website=Gizmodo}}</ref><ref>{{Cite web |last=Shugerman|first=Emily |date=1 Apr 2026 |title=Kids groups say they didn’t know OpenAI was behind their child safety coalition |url=https://sfstandard.com/2026/04/01/openai-ai-kids-safety-coalition/ |url-status=live |archive-url=https://ghostarchive.org/archive/g4oN8|archive-date=2026-04-04 |access-date=4 Apr 2026 |website=The San Francisco Standard}}</ref>

==Products==

@@ Line 26: / Line 26: @@
 ===Web Crawlers ignoring robots.txt (2025)===
-In 2025, Jonathan Bailey from PlagiarismToday posted an article going into how ChatGPTs web crawlers were ignoring the sites Robots.txt file.<ref>https://www.plagiarismtoday.com/2025/07/23/chatgpt-ignores-robots-txt-rehashes-my-column/ ([http://web.archive.org/web/20260106080839/https://www.plagiarismtoday.com/2025/07/23/chatgpt-ignores-robots-txt-rehashes-my-column/ Archived])</ref> PlaigarismToday had blocked OpenAI's web crawlers in August of 2023, yet the latest ChatGPT model at the time provided data from articles that were posted the day before on the website, even though OpenAI wasn't supposed to be scraping these web pages. This can be problematic for smaller websites, due to OpenAI's aggressive approach to web crawling, with their crawlers reportedly in a single week sending in more than 29 thousand requests to a wiki known as The Cutting Room Floor.
+In 2025, Jonathan Bailey from PlagiarismToday posted an article going into how ChatGPTs web crawlers were ignoring the sites Robots.txt file.<ref>https://www.plagiarismtoday.com/2025/07/23/chatgpt-ignores-robots-txt-rehashes-my-column/ ([http://web.archive.org/web/20260106080839/https://www.plagiarismtoday.com/2025/07/23/chatgpt-ignores-robots-txt-rehashes-my-column/ Archived])</ref> PlaigarismToday had blocked OpenAI's web crawlers in August of 2023, yet the latest ChatGPT model at the time provided data from articles that were posted the day before on the website, even though OpenAI wasn't supposed to be scraping those webpages. This can be problematic for smaller websites, due to OpenAI's aggressive approach to web crawling, with their crawlers reportedly in a single week sending in more than 29 thousand requests to a wiki known as The Cutting Room Floor.
 ===ChatGPT Atlas and prompt-injection vulnerability (2025)===
 In 2025, Brave posted an article about vulnerabilities that have agentic web browsers, such as ChatGPT Atlas, that consists of adding hidden malicious prompts in files, text or another media. Those prompts, combined with weak safeguards of the AI agents, can make them to expose and leak sensitive data of the user.<ref>https://owasp.org/www-community/attacks/PromptInjection ([http://web.archive.org/web/20260210124436/https://owasp.org/www-community/attacks/PromptInjection Archived])</ref>
+===Funding of the Parents & Kids Safe AI Act and creation of a child safety organization (2026)===
+In January 2026, OpenAI partnered with Common Sense Media and funded approximately $10 million dollars to support the California [[Parents & Kids Safe AI Act]] bill. The bill proposes to add [[age verification]], improvement of parental controls, and to prohibit [[Personalized ads|targeted advertising]] to underage users. On January 8, OpenAI created an organization, named Parents & Kids Safe AI Coalition.
+In March 2026, the OpenAI's organization sent emails to child safety coalitions, only mentioning it was sponsored by Common Sense Media. The purpose of these e-mails was to convince the child safety coalitions to support the bill.<ref>{{Cite web |last=Dellinger |first=AG |date=1 Apr 2026 |title=Group Pushing Age Verification Requirements for AI Turns Out to Be Sneakily Backed by OpenAI |url=https://gizmodo.com/group-pushing-age-verification-requirements-for-ai-turns-out-to-be-sneakily-backed-by-openai-2000741069 |url-status=live |archive-url=https://web.archive.org/web/20260401225238/https://gizmodo.com/group-pushing-age-verification-requirements-for-ai-turns-out-to-be-sneakily-backed-by-openai-2000741069 |archive-date=2026-04-01 |access-date=4 Apr 2026 |website=Gizmodo}}</ref><ref>{{Cite web |last=Shugerman|first=Emily |date=1 Apr 2026 |title=Kids groups say they didn’t know OpenAI was behind their child safety coalition |url=https://sfstandard.com/2026/04/01/openai-ai-kids-safety-coalition/ |url-status=live |archive-url=https://ghostarchive.org/archive/g4oN8|archive-date=2026-04-04 |access-date=4 Apr 2026 |website=The San Francisco Standard}}</ref>
 ==Products==