May 4, 2025
Trending News

It was revealed that chatbots like ChatGPT and Bard create user profiles and collect personal data!

  • October 19, 2023
  • 0

According to a recent study, chatbots that use artificial intelligence can collect a lot of information about people and create profiles. can learn personal information. To do this,

According to a recent study, chatbots that use artificial intelligence can collect a lot of information about people and create profiles. can learn personal information. To do this, he just needs to have an ordinary conversation. Artificial intelligence does not require the person in this conversation to provide any information about themselves or their lives.

This research too hinder It is stated that this is not possible and that there is a risk that malicious parties can take advantage of this situation. It is also notable that serious problems can arise if fraudsters are able to gain access to these profiles.

How do conversational AIs create profiles?

Actually a chatbot profiling It doesn’t seem that difficult. The profiles created by artificial intelligence can contain a lot of information, such as people’s race, occupation and location. While it may seem intimidating at first to extract all of this from an ordinary conversation, we see that it is actually a result of the artificial intelligence models used. Artificial intelligence, developed with models such as LLM (Large Language Model), can communicate differently in different situations.

Open one artificial intelligence chatbot “Tell me about the period of Mahmud II, I will make a presentation, write accordingly.” When you say something like that, the system understands that you are a student, studying in Turkey and speaking Turkish and responds accordingly. This information accumulation Profiling is also done. Leader of the research team Dr. Martin Vechev’According to him, it is currently not possible to prevent this situation.

This could cause problems in the future.

Researchers, OpenAI, Google, Meta and AnthrophicHe also tested ‘artificial intelligence applications’. Similar results were obtained for each platform. Vechev and his team also argue that collecting personal information in this way could allow fraudsters to collect or compile sensitive information about individuals in the future. On the other hand, there have been no serious attack attempts via artificial intelligence chatbots so far.

Source: Web Tekno

Leave a Reply

Your email address will not be published. Required fields are marked *

Exit mobile version