Google's Updated Privacy Policy
Google's updated privacy policy explicitly includes language about using 'publicly available' data to train its AI models, including Bard and Cloud AI capabilities. This move aligns with the ongoing trend of AI companies relying on user data to improve their algorithms. However, the legal and ethical implications of using content from sources like Wikipedia, news posts, and tweets without explicit permission from their creators remain contentious.
The updated privacy policy, which comes into effect on May 25, highlights the extent to which the search giant intends to collect and utilize user data for AI training purposes. The company's document notes that 'privately generated data' will continue to play a role in product development and AI enhancements. However, it also expands on the types of publicly available sources it can draw from to improve its algorithms.
'Google collects and uses many different types of data about you,' the document reads. 'We use this data to provide our services, to make products and services more useful, to provide customer support, and to share with others, as described in this Privacy Policy.' The document then lists a variety of types of data, including search history, watched videos, purchases, conversations, and location data from Android devices, and more.
But it also notes that Google will 'continue to get data from publicly accessible sources (such as newspaper articles, posts on public forums, and blogs) to help improve the accuracy of our AI systems, including those that are used by Google Translate, Bard, and our Cloud AI capabilities.'
Bard and ChatGPT
Bard is Google's alternative to ChatGPT, its embattled AI chatbot that caused an uproar last year when it began repeatedly swearing during conversations. While Bard has avoided the same sorts of gaffes, it has faced challenges, with some Google employees expressing concerns about its accuracy.
The broader use of publicly available data for training purposes certainly seems likely to improve Google's AI bots, but it also raises questions about how it intends to obtain and use that data. OpenAI, the creator of ChatGPT, is currently facing multiple lawsuits over its data harvesting practices.
The updated policy, similar to OpenAI's, certainly raises questions about whether content from sources like Wikipedia, news posts, and tweets can be freely used without violating the terms of their various licenses. Some authors, such as Mona Awad and Paul Tremblay, have even filed lawsuits against OpenAI, claiming copyright infringement over their online writings that were used to train ChatGPT's algorithms.
We've reached out to Google for further clarification on the policy changes and will update this piece accordingly. In the meantime, it serves as a timely reminder about the ways in which your data is used to improve AI algorithms, and all the potential privacy implications that may come with it.
Get all your questions about privacy answered in our 20 Q&A on privacy.
