Anthropic announced, on Thursday, that it will prevent its users from engaging in any “persistent and unjustified abusive or cruel behavior” toward its artificial intelligence program, Cloud, according to what was reported by Agence France-Presse.
The decision will be implemented as of November 12, and will address extreme cases “in which users repeatedly deal harshly with forms without a clear purpose,” according to what the company wrote in an update to the policy for using its tools.
She added that the decision will not include normal annoyance, differences of opinion, writing violent or dark fictional stories, or conducting tests and research on models.
The American company linked this decision to a feature it launched in August 2025, which allows “Cloud” to end a conversation on its own with a user who is constantly committing abuse. The company said at the time that this feature was the result of its “exploratory research into the psychological safety that artificial intelligence systems may have.”
Anthropic raises widespread controversy because it clearly addresses the possibility of its models possessing some form of consciousness.
Last February, its CEO, Dario Amodei, said in a New York Times podcast: “We don’t know if the models have consciousness,” while at the same time expressing his openness to this idea.
Spotify announced the expansion of its audiobook service from 22 markets to more than 180 markets by the end of 2026. In a move that includes broad markets in the Middle East, Africa, Asia, Europe, and the Americas, it gives more than 750 million people access to audiobooks via the platform.
The announced list includes several Arab markets, including Saudi Arabia, the Emirates, Qatar, Kuwait, Bahrain, Oman, Egypt, Jordan, Lebanon, Morocco, Tunisia, Algeria, Libya and Palestine, which gives the expansion a clear regional dimension after the service was previously mainly concentrated in a limited number of markets.
More than 350 thousand titles for new markets
Spotify said that users in the new markets will be able to access more than 350,000 titles, within a library covering more than 120 languages. The company works with more than 300 publishers around the world to expand the content available in different languages and markets.
This number differs from the total size of the audiobook library on the platform, as Spotify says in other data that its global catalog has exceeded 700,000 titles, while more recent reports linked to the Frankfurt Declaration indicated that the total number has now exceeded 800,000 audiobooks.
In Saudi Arabia, audiobooks were already available before this announcement under the “Premium” service, with 12 hours of listening per month for eligible subscribers and more than 350,000 titles within the subscription catalog, in addition to the ability to purchase individual titles via the web player. Official support data shows that Saudi Arabia and the UAE were among the 22 markets in which the service was available before the current global expansion.
New Arab markets
The new announcement means that the regional spread expands beyond Saudi Arabia and the UAE, to include the rest of the Gulf markets and countries in the Levant and North Africa. Spotify has not yet announced a unified model for accessing audiobooks in all of these markets. In some countries, the company provides 12 hours per month under the “Premium” subscription, while in other markets it relies more on the “Audiobooks+” service, which allows purchasing additional hours, in addition to the option to purchase a specific book directly and retain the ability to listen to it. The company confirmed that prices and access models will vary depending on the market.
Spotify says Audiobooks+ exceeded one million subscribers in just over one year, while the service's annual recurring revenue exceeded $100 million. The addition also allows subscribers to the “Family” and “Duo” plans to obtain additional hours of listening, with plans to expand it to other categories of subscribers.
Growth in listening
The expansion comes after the number of monthly audiobook listeners on Spotify increased by about 40 percent over the past year, while the total listening hours increased by more than 30 percent, according to figures announced by the company during the Frankfurt Book Fair.
Spotify had indicated earlier this year that nearly half of its audiobook users are under the age of 35, while listening hours increased by 60 percent between 2024 and 2025.
New tools for discovering books
Along with geographic expansion, the company is working on developing the in-app book experience. Among the tools it introduced is “Page Match,” which allows the user to switch between a paper or electronic book and its audio version, by matching the page to the listening position. Spotify says more than a million people use the feature monthly.
During 2026, the company will also launch tools to discover books through custom lists and filters, in addition to conversational search and creating playlists for audiobooks, in an attempt to integrate the book experience more deeply into the application that includes music and podcasts. The service will continue to be rolled out in batches until the end of 2026, with launch dates, subscription models, and prices varying from one market to another, while the Middle East markets represent an essential part of the largest geographic expansion of audiobooks on Spotify to date.
OpenAI revealed more details about the age prediction system within GBT Chat, which uses account-related signals and usage behavior to estimate whether a user may be under 18, rather than relying solely on the date of birth entered when creating the account. If the system believes the account belongs to an eligible teen, the “ChatGPT for Teens” experience is automatically activated with additional content controls and some feature limitations.
According to OpenAI, the age prediction system looks at a set of signals, including the general topics that the user discusses, the times that ChatGPT is used, how and when the account is used, activity patterns over time, in addition to the age of the account itself. This system does not replace the date of birth, but rather acts as an additional layer of security. Therefore, an account can be classified as a teen experience even if the user entered his date of birth when registering, if other indicators show that he may be under 18 years old. The company also acknowledges that the system may make errors in some cases.
What changes for users under 18?
When an account is labeled as potentially belonging to a teen user, ChatGPT can enable a personalized experience that handles some types of content more cautiously, such as explicit violence, challenges that may encourage risky behavior, sexual, emotional, or violent conversation roles, as well as content that promotes extreme beauty standards, unhealthy diets, or body image-related bullying.
The user does not need to activate the experience manually, as OpenAI says it is automatically applied to eligible accounts. The account remains active, with the ability to continue using “ChatGPT” for learning, creativity, and asking questions, while the company confirms that it does not display ads on teenagers’ accounts.
Educational tools and family controls
OpenAI has globally expanded the “ChatGPT for Teens” experience with a set of tools for daily use and study. Among them is (Study Mode), which guides the user through steps and questions instead of providing the answer directly, and can be suggested when using the platform for assignments or learning. The experience also includes features such as (Quiet Hours), which allows setting time periods to limit access to the service.
Parents or guardians can link their accounts to a teen's account via parental controls, allowing some settings to be managed. OpenAI explains that these controls do not give parents the right to view the teen's conversations themselves.
What happens if the system makes an error?
If a user is 18 or older and is mistakenly placed in the teen experience, they can review the age classification (Settings > Account > Age), then use the age verification option if it is available in their account.
OpenAI relies on a third-party company called Persona to perform the verification process. Depending on the country, the user may be asked to provide a live facial image, a government identification document, or both. The company says that “Persona” deletes the ID photo or face image within seven days of verification, while “OpenAI” does not receive a copy of the ID or photo itself, but rather the information necessary to determine the appropriate experience for the account. If it is proven that the user is over 18 years old, the restrictions for teenagers are removed. If the verification process shows that the account should remain within the teen experience, protection may continue or be restored to the account.
An extra layer of security
OpenAI explains that age prediction is not a final system for judging a user's age, but rather an additional mechanism that helps it apply the appropriate experience to younger accounts. If a user chooses to verify their age via Persona, the company says it will no longer continue to run the age prediction system on that account. With this approach, OpenAI combines date of birth, usage signals, and external age verification to determine when controls for teens should be implemented within GBT Chat.
