A study on privacy of AI chatbots and session hijack

Krithiga, B and Marikkannan, M (2025) A study on privacy of AI chatbots and session hijack. International Journal of Science and Research Archive, 15 (3). 046-054. ISSN 2582-8185

The paper revolves around our daily usage of AI(s) and the rise of session hijack attackers on the web. Whatever people surf on the web is potentially vulnerable to attackers because the network is a web of connections, and the attackers are always finding a way to attack users and steal their data. Nowadays, using AI has become a part of daily routine regardless of field, profession, or usage. Most people use AI on the web since it is the most convenient way of using it without installing it on the device. However, most people don’t know that using something on the web may lead to session hijacking. So, people want to find a way to do their daily routines with AI without becoming victims of session hijack. As someone who codes from time to time, people may have some knowledge of the Integrated Development Environment (IDE) and environment data. Using something in an IDE has no connection to the internet by any means, even while connected to an internet service, because an IDE is an isolated environment built into the system for coding purposes while also allowing us to connect with the network. This basically works by avoiding things like spiders, cookies from the web, which will be stored if a user uses the browser, but this has no effect on IDEs. Running AI on localhost is also a possibility, but it demands high specifications on the hardware, such as a large amount of RAM, VRAM and external GPU.