On July 1, 2023, Google announced an update to its privacy policy. The tech giant revealed its intention to utilize publicly available data obtained through web scraping for training its artificial intelligence (AI) models.
The revised privacy policy explicitly states that Google will leverage this information to improve its services, as well as to develop new products, features, and technologies that ultimately benefit users and the wider public.
Language translation stands among the areas where AI models will be trained using this data. According to the new policy, Google “may collect information that’s publicly available online or from other public sources to help train Google’s AI models and build products and features like Google Translate, Bard, and Cloud AI capabilities.”
To improve its AI models and services, Google utilizes various technologies to process user information. Automated systems are employed to analyze user content enabling the delivery of personalized features such as customized search results and tailored advertisements based on individual preferences. This data processing also aids in the detection and prevention of abusive content, including spam, malware, and illegal materials.
Google updates its privacy policy to say the company may use “publicly available information” to train its AI models and build products like Translate and Bard (@thomasgermain / Gizmodo)https://t.co/PmbSKPg5Fuhttps://t.co/F263krViQo
— Techmeme (@Techmeme) July 4, 2023
Furthermore, Google relies on algorithms to recognize patterns in data, enabling services like Google Translate to facilitate cross-language communication by detecting common language patterns in the phrases users input for translation. Through these algorithms, Google aims to improve the accuracy and effectiveness of its language-related AI models.
Shifting the Focus
One of the notable differences between the updated policy and the previous version is the expanded scope of Google’s AI product lineup.
The previous version emphasized the use of public data for updating and training language models, while the updated policy emphasizes the utilization of all public data to update AI models.
More specifically, before the update, the policy section specified the use of data for “language models” rather than “AI models.” Additionally, it solely mentioned Google Translate, while the updated policy includes Bard — Google’s large language model (LLM) — and Cloud AI as well.
Predictably, the policy change has garnered attention and raised concerns about privacy and data protection.
Over the long weekend, Google updated its privacy policy to collect public data for AI training.
The data will train Google’s AI models and build products/features such as Google Translate, Bard, and Cloud AI capabilities.
Key points:
-Google can use any information people post… pic.twitter.com/JOhmA5rBHY
— Rowan Cheung (@rowancheung) July 4, 2023
In an interview with The Verge, Google spokesperson Christa Muldoon clarified that the change aims to provide transparency, clarifying that new services like Bard are included in the practice of training AI models using publicly accessible information.
“Our privacy policy has long been transparent that Google uses publicly available information from the open web to train language models for services like Google Translate,” she said. “This latest update simply clarifies that newer services like Bard are also included. We incorporate privacy principles and safeguards into the development of our AI technologies, in line with our AI Principles.”
