In a recent update to its privacy policy, Google has announced that its AI services, including Bard and Cloud AI, may now be trained on public data collected from the web. The revised policy effective July 1, 2023, says the tech giant can scrape public data on the web.

The move has drawn attention to the company's practices and raised concerns regarding privacy and the use of copyrighted materials. 

The updated policy clarifies that Google uses publicly available information to train language models for services like Google Translate, with newer services like Bard also included. Google emphasizes its commitment to privacy principles and safeguards aligned with its AI Principles.

Google will utilize the information to enhance its services and develop new products and features for the benefit of users and the public. It acknowledges the use of publicly available information to train Google's AI models, including Google Translate, Bard, and Cloud AI capabilities.

Scrape public data including copyrighted info

While the update provides additional clarity on the services trained using collected data, it leaves unanswered questions about how Google prevents copyrighted materials from being included in the data pool.

 Some publicly accessible websites have policies prohibiting data collection or web scraping for training AI models, prompting concerns about the potential misuse of data without explicit consent, particularly under regulations such as GDPR.

The legal implications of using public data for training AI systems remain unclear. The ambiguity surrounding the fair use doctrine in this context has sparked lawsuits and calls for stricter regulations governing AI companies' data collection and usage. Additionally, there are concerns about ensuring the responsible processing of data to prevent dangerous failures within AI systems. 

Those responsible for sifting through vast pools of training data often face grueling work hours and extreme conditions.

Gannett, the largest newspaper publisher in the United States, has filed a lawsuit against Google and its parent company, Alphabet, alleging that the search giant's advancements in AI technology have granted it a monopoly over the digital ad market. 

Google's AI search beta has been criticized as a "plagiarism engine" and accused of depriving websites of traffic.

Meanwhile, popular social platforms Twitter and Reddit, which host substantial amounts of public information, have implemented significant measures to prevent unrestricted data harvesting by other companies. Changes to their APIs and limitations on data access have faced backlash from their respective user communities due to the negative impact on core user experiences.

As the use of public data for training AI models becomes more prevalent, debates around privacy, data ownership, and responsible usage are likely to intensify. Striking a balance between innovation and protecting user rights will continue to be a complex challenge for tech companies and lawmakers.