Sunday, August 17, 2025
HomeTechnology"Incognito modes": can we really believe AI?

“Incognito modes”: can we really believe AI?

"incognito modes": can we really: This article explores the topic in depth.

However,

"incognito modes": can we really:

Mode incognito? However, Image: ChatGPT. "incognito modes": can we really
Mode incognito? Nevertheless, Image: ChatGPT.

The tools based on artificial intelligence (AI). such as Chatgpt, Gemini or Perplexity, now offer temporary modes allowing to dialogue without the exchanges being kept in the long term. This advance aims to reassure the user anxious to preserve their privacy, while responding to increasing concerns concerning data protection.

With the introduction of his new “temporary discussion” mode. Gemini guarantees that conversations will not be used to personalize the service or to perfect AI models. “incognito modes”: can we really Chatgpt. with his ephemeral cat, and Perplexity, with his incognito mode, adopt a comparable approach: shared information remains outside the usual historical and theoretically serves to train the models.

However, the user remains dependent on the shelf life chosen by each actor. For example. Google specifies that the data is kept for 72 hours by Gemini, in particular to ensure the safety and consistency of the experience, before being deleted, according to information in connection above. Perplexity articulates a duration of 24 hours, while Openai reserves the right to keep this information 30 days in Chatgpt!

Respect for privacy: between speech. reality

Theoretically, the Internet user controls his activity and can activate or deactivate exchanges in the settings. However. it is essential to understand exactly what “activity data” or “televersation” covers: these are the files, images or videos transmitted to the tool, which can be analyzed to “improve the service” “incognito modes”: can we really if the recording option is active.

The dissociation of account conversations, before any treatment with third parties, is supposed to strengthen confidentiality. But the passage through “human examiners” within companies to verify the content can raise the question of the real. anonymization of the information. Too often, there is a certain blur in the matter. Many questions remain unanswered …

Should you trust?

Finally, everything is based on the good faith and the ethics of societies developing and exploiting these AI tools. None of them is subject to systematic control by an independent entity ensuring compliance at all levels of the. promises issued. As I have already written, it clearly lacks regulation in the matter.

The citizen concerned with his private life must therefore wonder and remain vigilant. Without the implementation of an outdoor. completely independent audit system, there is no guarantee that internal practices comply with the commitments displayed. “incognito modes”: can we really Caution remains in order to share sensitive information, even in “temporary”, “ephemeral” or “incognito” modes.

Xavier Studer


Learn more about the High-Tech & Telecom blog by Xavier Studer

Subscribe to get the latest posts sent to your email.

Further reading: A new snapchat function to reassure your friends at the end of the eveningQualcomm Snapdragon: Adobe launches a native version of Premiere Pro for ARM laptopsHonor, Xiaomi, Samsung: discover this selection of 3 high -performance smartphones at less than 100 eurosThe future of spatial communication: ESA succeeds in a laser connection with a probe at 265 million kilometers-45 % on this connected watch very popular with athletes.

reagan.west
reagan.west
Reagan live-tweets NASA launches and follows up with long-form explainers that replace jargon with playground metaphors.
Facebook
Twitter
Instagram
RELATED ARTICLES

LEAVE A REPLY

Please enter your comment!
Please enter your name here

- Advertisment -

Most Popular

Recent Comments