The Dutch Data Protection Authority (AP) has expressed its concern regarding the manner in which open-source Artificial Intelligence (AI) company, OpenAI, is handling personal data. This apprehension was prompted by the introduction of a chat application – ChatGPT – to the Netherlands approximately four months ago; since then, it has been accessed by 1.5 million people.
Powered by an advanced language model, ChatGPT responds to posed questions with perceived accuracy. Consequently, this necessitates the collection and storage of data during the program’s training stages. The AP accordingly seeks clarification on the process through which such data acquisition is achieved, particularly with regards to the collation of questions from its users and compilation of material already existing online. Furthermore, they have highlighted the potential for the generated content to be misleading, potentially offensive or even wrong. Insofar as rectifying or eradicating such content, there are still underlying uncertainties as to what can be done.
In light of the foregoing, the Dutch Data Protection Authority has deemed it crucial to conduct a thorough investigation to ascertain whether the use of ChatGPT abides strictly with current privacy laws and regulations.