Artificial Intelligence Search Engine Previews Users’ Privacy Concerns
OpenAI has provided the first glimpse of its new AI-powered search tool called SearchGPT. While the prototype looks promising for answering queries directly from the web, it begs the question – what user data will be collected and how protected is privacy?
Currently in limited alpha testing with 10,000 users, SearchGPT appears capable of responding instantly to all manner of questions from “what’s the weather today?” to “upcoming music festivals”. OpenAI aims to integrate some of its features into the popular ChatGPT bot as well.
When it comes to data handling, general location from IP addresses may be anonymously shared for more localized results. Searches themselves could also be stripped of personal identifiers and shared externally. Reassuringly, OpenAI pledges to keep conversations within its AI services like SearchGPT and ChatGPT separate.
Privacy controls are available to users. Location tracking can be disabled, as can personalized result improvements. And while IP addresses still provide a general area, more precise coordinates are optional. Search histories may also be deleted from OpenAI systems within 30 days.
However, some data cleansing like anonymization remains outside of deletions. And past issues like ChatGPT’s plaintext storage on devices cause valid concerns. Moving forward, increased transparency on exactly what data is used, shared, secured and removable will be key to building trust in AI applications that handle our most private thoughts and questions on a daily basis.
