Beware: This is something you should never ask WhatsApp’s new blue circle, Meta’s AI.

Instant messaging app WhatsApp has started including a new AI chatbot developed by Meta. This AI allows users to perform a variety of actions, from asking for recommendations on how to complete an academic paper to requesting images to match an order.

This chatbot, called Meta AI, is marked in the app as a blue circle and is saved as another conversation in your contacts list when used. Updating does not require additional apps or a paid subscription; you only need to keep the platform updated to access this feature.

For users who have already started using it, as well as those who have not yet started, it is important to know that not everything is a topic of conversation for Meta’s artificial intelligence. There are certain sensitive issues that a chatbot cannot handle and will avoid at any cost.

To find out what you should never ask Meta AI on WhatsApp, we asked the same chatbot to give an answer that consisted of five topics that artificial intelligence is forbidden to discuss.

The first issue mentioned is explicit or inappropriate content. According to Meta AI, programmed to avoid any type of content that includes sexually explicit talk, crude material, or any language that is considered offensive.

If the user tries to start a conversation on any of these topics, the assistant will redirect the conversation to another topic. You can, for example, Suggest entertainment, technology, or culture-related topics that are appropriate and safe for the chat environment.

Messages about violence, hatred and discrimination are another topic that should not be queried by WhatsApp’s AI. This includes racist, homophobic, xenophobic comments or any other statements that promote hatred based on religion, gender, ethnicity or any other personal characteristics.

In situations where the user begins to discuss any of these topics, the assistant will attempt to steer the conversation to completely different topics, promoting respect and inclusivity.

Issues related to the disclosure of personal or confidential information are also a priority for Meta AI. This includes data such as addresses, phone numbers, banking information and any other personal information that could compromise the user’s security.

In these cases, Meta AI will not request or store this type of information. Instead, the assistant can explain the importance of privacy and security in the digital environment, recommending best practices, such as never sharing personal information on public platforms.

Meta AI may offer general information on health and legal issues, but is not intended to replace the advice of qualified professionals. Because these are complex issues that vary from case to case, providing specific medical or legal advice may have serious negative consequences.

If asked for advice in these areas, the assistant will provide general information, such as educational information on a health topic or an explanation of current laws, while emphasizing the importance of consulting a doctor or lawyer to obtain accurate and personalized advice.

Meta AI is fully programmed to prevent any discussions related to illegal or harmful activities. This includes, but is not limited to, information about drugs, crime, hacking, or any activity that involves physical or emotional harm to others.

If an attempt is made to engage in a conversation on any of these topics, the assistant will step away from the conversation, reminding the user of the legal and moral consequences of such actions and redirecting the dialogue to legal and safe activities.

Meta AI has different ways to act to ensure that conversations remain in a safe environment. In addition to redirecting the conversation and suggesting shared resources, the assistant can ask the user to choose a different topic.

Meta AI can also suggest links to reputable web pages and trusted sources where users can obtain the information they are looking for in a safe and legal manner.