Danish DPA Warns Against Use of AI Tools for Filing Complaints
Be careful when using AI to generate privacy complaints and inquiries, the Danish Data Protection Authority cautioned Monday.
Sign up for a free preview to unlock the rest of this article
Privacy Daily provides accurate coverage of newsworthy developments in data protection legislation, regulation, litigation, and enforcement for privacy professionals responsible for ensuring effective organizational data privacy compliance.
AI may help some people, but it carries risks, the watchdog said. One is that AI tools can share the complainant's and others' personal data unintentionally because of how they work. They often function in such a way that the user's information is processed to deliver a service but also used for the AI provider's purposes, such as to further develop the tool.
That means that without intending to, people filing complaints can end up sharing health information, social security numbers, information about their kids and other personal data with the AI provider, the DPA said. That data can then become part of the AI tool's data set and be included in responses to other users.
Often, "it will be practically impossible or very difficult to gain control over how one's information is subsequently used."
Additionally, the DPA noted, AI-generated complaints tend to distort the content of a complaint by making mistakes and unnecessarily complicating the facts and associated laws. The tools are only as correct as the information they have access to. If a tool's database isn't complete or up-to-date, its answer may look convincing and correct, but not be.
The DPA also noted that some complainants use AI tools to predict the outcome of their case, adding that AI can't predict the Danish DPA's decisions.
The authority recommended that complainants formulate complaints in their own words, noting that if the DPA has questions about the complaint, it will contact the complainant by phone or in writing.
If someone insists on using an AI tool to make an appointment with the DPA, it said, don't enter confidential and sensitive personal data and other information you don't want shared and disclosed with the AI tool provider.
Despite the fact that many providers allow users to opt out of collecting information for further model training, users should be aware that in some cases, a provider might collect, process and disclose information about personal matters.