Google will use public data to train its AI models as part of its new privacy policy

Priya Walia

deepmind

Looking for more info on AI, Bing Chat, Chat GPT, or Microsoft's Copilots? Check out our AI / Copilot page for the latest builds from all the channels, information on the program, links, and more!

Google has recently made adjustments to its privacy policy to specify that publicly accessible data may be utilized in the training of its AI models. To be more precise, the wording of the policy has been modified to replace “AI models” with “language models.”

Furthermore, the company has emphasized that it has the potential to utilize publicly available information not only to create features but also to develop complete products such as Google Translate, Bard, and Cloud AI capabilities.

The policy update aims to inform and clarify to individuals that any content they publicly post online could be employed in training Bard, its future iterations, or any other generative AI product brought forth by Google.

The tech giant has publicly disclosed these modifications to its privacy policy through its archive.

Also, the previous wording of the policy indicated that Google utilized the data solely for constructing “language models” for its Translate service. However, the revised policy now encompasses Bard and introduces “Cloud AI.”

This implies that any content shared through a Google product such as YouTube, Gmail, or Search will be retained by the prominent American technology corporation to train its artificial intelligence systems.

Google recently announced nine notable artificial intelligence (AI) features during the Google I/O event. However, several organizations expressed concerns about the source of information for the large language models (LLMs) that power these projects.

The situation for AI companies is becoming more challenging as the European Union (EU) prepares to implement stringent regulations on AI products. To comply with these regulations, Google had to halt the launch of its Bard AI in Europe last month. Furthermore, popular platforms such as Reddit, Twitter, and others have declined to offer unrestricted access to their application programming interfaces (APIs).

These APIs are crucial for obtaining extensive post archives and potentially utilizing the content for AI training. Meanwhile, ChatGPT, Google’s prominent competitor in the AI market, has recently implemented data controls for users worldwide following a complaint from Italy’s data protection authorities.

Via Tweak Town