¿Se puede desactivar la Inteligencia Artificial de WhatsApp? – El Sol de Puebla

Meta has integrated Meta AI, a generative artificial intelligence assistant, across WhatsApp, Instagram, and Messenger. According to official Meta product documentation and company announcements, users can interact with the AI to ask questions, generate images, and seek information directly within their chat interfaces without leaving the apps.

The rollout allows users to trigger the assistant by typing @Meta AI in a group chat or by starting a direct conversation with the bot. Meta states that the AI is powered by Llama 3, the company’s latest large language model, which is designed to handle more complex reasoning and provide more accurate responses than previous iterations.

Users can utilize the tool for tasks such as planning trips, drafting messages, or creating AI-generated images using the “Imagine” feature. Meta has integrated the assistant into the search bars of these platforms, replacing or augmenting traditional keyword searches with conversational AI responses.

Regarding the ability to disable the feature, reporting from El Sol de Puebla indicates that Meta AI is integrated into the core infrastructure of the messaging apps. While users can delete the specific chat thread with Meta AI or avoid interacting with the bot, there is no global “off” switch in the settings menu to completely remove the AI functionality from the app’s interface.

This integration follows a broader industry trend of embedding LLMs (Large Language Models) into existing social ecosystems. By placing the AI in WhatsApp and Messenger, Meta targets a user base of billions, attempting to compete with standalone AI agents like OpenAI’s ChatGPT and Google’s Gemini.

Meta has addressed privacy concerns by stating that personal messages and calls remain end-to-end encrypted and are not used to train the AI models. However, the company notes that prompts sent directly to Meta AI are used to improve the assistant’s performance, though these are distinct from private user-to-user conversations.

The deployment of Meta AI is not uniform globally. Due to varying regional regulations, particularly the European Union’s General Data Protection Regulation (GDPR) and the AI Act, certain features have been delayed or modified in European markets. Meta has adjusted the rollout in these regions to comply with local data processing and privacy laws.

Technical specifics of the Llama 3 integration include improved multilingual support and a reduction in “hallucinations”—instances where AI generates false information. Meta claims that the model’s training data is more diverse, allowing it to understand cultural nuances better than the Llama 2 version.

For users seeking to manage their interaction with the tool, the current options are limited to:

• Deleting the Meta AI chat thread to remove it from the primary chat list.
• Ignoring the AI suggestion in the search bar.
• Avoiding the use of the @Meta AI tag in group conversations.

The inability to fully deactivate the AI has sparked discussions regarding user agency and “dark patterns” in software design, where features are added that cannot be easily removed. Despite this, Meta continues to push the integration as a primary method for users to access information without switching applications.