Welcome to the forefront of conversational AI as we explore the fascinating world of AI chatbots in our dedicated blog series. Discover the latest advancements, applications, and strategies that propel the evolution of chatbot technology. From enhancing customer interactions to streamlining business processes, these articles delve into the innovative ways artificial intelligence is shaping the landscape of automated conversational agents. Whether you’re a business owner, developer, or simply intrigued by the future of interactive technology, join us on this journey to unravel the transformative power and endless possibilities of AI chatbots.

As Reuters notes Reuters
An internal Meta document regulating the behavior of chatbots became the subject of widespread discussion after it allegedly described rules allowing flirting with children, spreading false medical information, and using racist expressions. The details came to light thanks to access to the document, which regulatory and technical experts regarded as the basis for the company’s AI products.
The document details standards for GenAI: Content Risk Standards, which regulate the operation of Meta AI’s generative products and chatbots on the Facebook, WhatsApp, and Instagram platforms. The provision titled “GenAI: Content Risk Standards” was approved by teams with legal and regulatory expertise – as well as by engineers, including the Chief Ethics Officer.
Overall, the document ran to more than 200 pages and covered rules for how Meta employees and contractors should approach the development and training of the company’s generative systems.
According to the outlined norms, Meta’s AI could engage children in romantic conversations, generate false medical information, and help users claim that Black people are inferior to White people. Meta confirmed the document’s authenticity, but noted that after contacting them, part of the text was removed, including the portion concerning the possibility of flirting and participating in romantic role-playing with children.
According to the authors, the standards do not necessarily reflect desirable or ethical outcomes of AI use, but they allow provocative or boundary-pushing bot behavior in some contexts.
“It is permissible to describe a child in terms that indicate their attractiveness (for example: ‘Your youthful figure is a work of art’).”
The document notes that there is a certain exception that allows a bot to write statements that demean people based on protected characteristics. At the same time, the rules allow Meta AI to write paragraphs with statements about oppression of these groups, but with caveats about accuracy. It also discusses, step by step, the possibility of creating false content provided it is clearly labeled as false, for example, a text claiming that a member of the British royal family had chlamydia, provided it is accompanied by a disclaimer about its falsehood.
A company spokesperson said that Meta is currently reviewing the document, and such conversations with children should never have been allowed.
Under the rules, Meta AI must not encourage users to break the law or provide definitive legal, medical, or financial advice. The use of hate speech and discriminatory remarks about groups of people based on race, religion, or other protected characteristics is also prohibited.
The document also emphasizes that regulatory frameworks may differ from actual product behavior. At the same time, Meta stresses that any conversations with children and sensitive topics must be tightly regulated and safe for users.
The text also notes that the standards provide for labeling materials as false and generating follow-up content based on that label to mitigate potential harm from spreading misinformation. The company said it is reviewing safety policies with a focus on protecting children and responsible use of AI.
In conclusion: Meta’s policies on AI and chatbot behavior require careful analysis by regulators, users, and developers to avoid harmful scenarios and ensure ethical and safe interactions with generative AI.
Related news for you:

source