La Inteligencia Artifical más privada que puedes usar en 2026

Summary

This video explores the privacy levels of popular AI tools, ranking six major platforms based on their data collection, retention, and sharing practices. From ChatGPT to Gemini and DeepSick, the video reveals how each AI handles user data and the risks involved. It offers practical advice on balancing privacy and usability, recommending specific tools for different needs—such as Cloud and Proton’s Lumo for enhanced security. Viewers learn how to safeguard their data while interacting with AI and avoid privacy pitfalls, with insights on what to share and how to adjust settings for better control.

Takeaways

• 😀 Privacy rankings of AI tools: The script ranks 6 popular AI tools based on privacy, from 6 (worst) to 1 (best).
• 😀 Data collection practices: AI tools collect various types of personal data, including IP, device identifiers, browsing history, and even typing patterns.
• 😀 Impact of feedback: Giving feedback to AI tools (e.g., thumbs up/down) can contribute to data collection, even if the user has disabled model training.
• 😀 ChatGPT privacy: ChatGPT collects 10 types of data and uses them to train its models. Free users can't disable data collection for model training.
• 😀 ChatGPT data retention: Even after deleting chats, data may be stored for up to 30 days for abuse monitoring, and potentially indefinitely with a court order.
• 😀 Pop-up ads in ChatGPT: Ads will become more personalized and potentially intrusive due to the data ChatGPT collects.
• 😀 DeepSic AI: Despite being open source, DeepSic collects 11 types of data and stores them in China, where data is subject to China's Intelligence Law.
• 😀 Copilot by Microsoft: Copilot collects data for advertising and model training unless the user opts for a paid plan. Conversations are kept for up to 18 months by default.
• 😀 Privacy features of Cloud AI: Cloud AI offers strong privacy controls, including an option to deactivate model training and limit data retention to 30 days.
• 😀 Gemini by Google: Gemini collects 22 types of data, including location, contacts, and browsing history, and stores data for up to 3 years under certain conditions.

Q & A

What factors were considered when analyzing the privacy of AI tools?

The analysis considered five main factors: personal data collection, how the models are trained using user interactions, whether the AI shares data with third parties, how long data is retained, and where the data is stored (which affects access by governments).

Which AI tool is considered the best for privacy?

Lumo by Proton is considered the most private AI tool because it encrypts conversations with the user's personal key, does not store records, and operates under Swiss privacy laws, which are outside of US jurisdiction.

What is the main privacy concern with Gemini by Google?

Gemini collects a large amount of personal data (22 out of 35 types) and integrates with various Google services, enabling detailed profiling. Despite having privacy controls, its data retention policies and access to all linked apps make it one of the least private AI tools.

How does Chat GPT handle user data?

Chat GPT collects up to 10 types of data, including IP address, device type, and user conversations. In the free version, data is used to train models, while the paid version offers options to control how data is used. Conversations are stored for 30 days, and can be retained longer under certain conditions.

What privacy measures does Cloud AI offer?

Cloud AI offers a strong balance between privacy and usability. Users can deactivate the training feature, use incognito mode for sensitive conversations, and retain data for only 30 days. It also provides a privacy rating of low risk and strong security controls.

Why is DeepSick's privacy considered controversial?

DeepSick is controversial because it stores data in China, where the government can access data without judicial orders under the National Intelligence Law, raising concerns about privacy and the potential for unauthorized access.

What is the difference in privacy practices between the free and paid versions of Chat GPT?

The free version of Chat GPT collects and uses all user data to improve its models, while the paid version offers more privacy options, such as disabling the use of data for training. However, some interactions, like feedback, may still be used for training even if data usage is disabled.

How does Microsoft Copilot handle user data for advertising?

Microsoft Copilot tracks user data for advertising and product improvements unless the user opts for a paid version. Even with the paid version, Microsoft retains conversations for up to 18 months by default, although privacy settings can be adjusted.

What is the main advantage of using DAC.AI for privacy?

DAC.AI, integrated with DuckDuckGo, ensures full anonymity by not requiring users to create an account or provide any personal information. It serves as an intermediary that removes metadata, protecting user identity during interactions.

What recommendation does the speaker make for users who prioritize privacy over usability?

For maximum privacy, the speaker recommends using Lumo or DAC.AI, especially for sensitive tasks. Users seeking a balance between privacy and usability should opt for Cloud AI with training disabled and use incognito mode for sensitive conversations.