AI in every aspect of life, from phone and apps to search engines and even DriveFor some reason. The fact is that we are now receiving web browser with baked-in AI assistants and chatbots, showing that the way some people are using the Internet and today are also very different to consume information a few years ago.
But the AI devices are more and more asking for the gross level of access to your personal data, which is under the guise of the need to work. Such access is not normal, nor should it be normalized.
Not long ago, you would be right to question that the free -looking free “flashlight” or “calculator” app will try to request access to your contacts, photos and even your real -time data. These apps may not require that data to function, but they will request it if they feel that they can make a reindeer or two by mudling your data.
These days, AI is not all different.
Take the latest AI-managed web browser of perplexity as an example, comet. The comate allows users to find answers with its underlying AI search engine and automatically automatches regular tasks such as email and calendar phenomena.
In recent hands with browser, Techcrunch found that when perplexity requests the user to reach the Google calendar, the browser asks for a wide swatth of permissions in the user’s Google account, including the ability to manage the draft and take a copy of a copy on all of his calendars.
Perplexity says that this data is stored locally on your device, but you are still providing the rights of the company Use and use Your personal information, which involves improving your AI model for all.
Perplexity is not alone in asking for access to your data. There is a tendency of AI apps that promises to save you time by transcript your call or work meeting, but which requires AI accessories to use your real-time personal conversations, your calendar, contacts and more. Meta, also, is testing the range of its AI apps, what can ask for access, including tapping in photos stored in the user’s camera roll that has not yet been uploaded.
Signal President Meredith Whitekar recently compared the use of AI agents and assistants to “put his brain in a jar”. Whittaker describes how some AI products can promise to do all types of worldly works, such as submitting a table in a restaurant or booking tickets for a concert. But to do this, AI will say that your browser is required to open to load the website (which can allow AI access to your stored password, bookmark and your browsing history), a credit card to create reservation, to mark your calendar, and it can also ask to open your contacts so that you can share with a friend.
There are serious security and privacy risk associated with using AI assistants that rely on your data. In allowing access, you are immediately and irreversibly handing over an entire snapshot of your most personal information as that time dating from your inboxes, messages and calendar entries in previous years, and more. All this is to do a task that saves you time – or, on the talk of Whitekar, prevents you from actively thinking about it.
You are also allowing AI agent to act autonomously on your behalf, which requires you to have huge amounts of faith in a technique that is already prone to obtain things wrong or flat. To use AI you need to rely on companies that develop these AI products, which rely on your data to try to perform your AI model better. When things go wrong (and they do a lot), it is common for humans in AI companies to find out why things do not work to find out your personal signals.
From the point of view of security and privacy, a simple cost-profit analysis of connecting AI to its most personal data is not just worth giving access to your most personal information. Any AI apps are asking for permission for these levels, your alarm bells should play, such as the flashlight app wants to know its location at any time.
Given the remarks of the data handed over to AI companies, ask yourself if you get out of it, it is really worth it.
