Generative artificial intelligence ( AI ) needs a lot of data to feed its models. Like Google! the most transparent publishers clearly announce their colors and let the user choose certain options…
On the side of Gemini! Google’s AI! we can read this warning! reproduced below in italics
Google’s expert reviewers
Just by chatting with Gemini! you help improve Google services! including the machine learning models that power Gemini.
This improvement process also requires specialized reviewers to handle your conversations.
So when using Gemini! don’t enter any information that you wouldn’t want a reviewer to have access to or Google to use.
Your Google Workspace content! like content from Gmail or Drive! is not reviewed and is not used to improve Gemini.
You can turn off Gemini App Activity
If you don’t want your future conversations to be reviewed or used to improve machine learning models! turn off Gemini App Activity. Google also has a privacy guide .
The first link above takes you to some settings in your Google line database account. It is then possible to disable the recording of activities in Gemini. You can even clear the history of requests already made.
On the ChatGPT side! a similar option exists. Here is in italics what is written in the settings in the “data” section:
Save new discussions on this browser to your history and allow them to be used to improve our models. Unsaved discussions will be deleted from our systems within 30 days. This setting does not sync across browsers or devices. Learn more .
Once again! Perplexity shows the way. Here too in the settings of your account it is possible to manage the use of your data! but apparently without any compensation for archiving over time… The warning text in italics:
AI Data Usage. This feature allows us to use your search data to improve our AI models. Disable this option if you want to exclude your data from this process. All info.
Finally! let’s take advantage
This short text to emphasize technology in the workplace that in the settings of most of these generative AIs it is possible to inform them about their profile! tastes and expectations to obtain more personalized results. Perplexity thus asks for multiple information on their activity! interests! etc. Obtaining more precise results quickly becomes intrusive. Perhaps a necessity…
As I have always said in this Beware of ChatGPT column: caution! caution cyprus business directory and caution with these generative AIs. Not only do they remain buggy to the core and produce many hallucinations! as Google has always pointed out! but they are also real data pumps. A forewarned user is always worth two.
