Google changed its Privacy Policy in favour of AI models

You are currently viewing Google changed its Privacy Policy in favour of AI models
Photo by Mitchell Luo on Unsplash

It is not a secret that practically nobody reads privacy policies – and Google counted on it. In their latest privacy policy update, Google explicitly states that it now reserves the right to curate vast amounts of publicly available online content to enhance its AI tools. What does it mean in practice? Well, if you’ve posted it, Google now owns it and it might just read it, integrate it, and use it to train its advanced chatbot technology.

Google’s Privacy Policy update takes a bold stance on an already legally controversial matter

In previous versions of Google’s privacy policy, tech giant declared that it used public internet sources to train large language models, such as Google Translate. With the latest privacy policy update, Google has taken a bold step forward. While the change from “language models” to “AI models” may seem minor, it means that Google now explicitly acknowledges its practice of internet scrapping for the refinement of its AI tools – such as Bard, Cloud AI, and previously mentioned Google Translate.

However, the language in this privacy policy update suggests that Google believes any content accessible on the public internet is considered fair game for leveraging to improve its products. This wouldn’t be as problematic as it is, if Google wasn’t the owner of services such as Google Search, Gmail or YouTube (which most of us internet users use on a daily basis). This, therefore, means that whatever users share through one of Google’s services, Google will save and use for training its AI models.

No surprises here – Google, the indisputable gatekeeper to the internet, now sends a clear message: there are no boundaries when it comes to leveraging public internet content to fuel its AI initiatives. Every morsel of content, from articles to posts and beyond (including the content of our emails), is now on the menu for Google’s AI algorithms training. Needless to say, the potential for innovation and advancements in AI-powered products is immense, but it comes with a set of challenges and ethical considerations.

Ongoing “right to scrape” debate

The tech community finds itself in the midst of a heated discussion regarding internet scraping. Besides the controversial Google privacy policy update, the recent issues faced by Reddit and Twitter, involving access to APIs and scraping, have added fuel to the fire. The impact of such practices on users and platforms is a topic that demands thoughtful examination and discourse. With Elon Musk, Twitter’s owner, vocal about scraping concerns, the debate has reached new heights.

Screenshot of Elon Musk’s tweet, 1st July 2023

OpenAI, the brains behind ChatGPT, also finds itself entangled in legal battles amid allegations of unlawful data collection from internet users and subsequent use of the information to develop its products.

Additionally, OpenAI is grappling with another lawsuit that revolves around copyright infringement and privacy violations. The claims suggest that the company employed copyrighted books without obtaining proper permission, using the texts to train its AI systems. In doing so, OpenAI allegedly failed to credit or compensate the copyright holders, leading to a contentious legal situation.

Legal challenges along these lines are broadly expected over the next few years as AI models and products continue to roll out, and some are in courts already – such as the above-mentioned case against OpenAI. The outcome of this case will definitely have significant implications for Google’s own plans in this domain, and wording in their privacy policy would most probably not be sufficient.

Taking into consideration Google’s dominance in various sectors of the tech industry (for which Google is also facing legal and regulatory threats), it might be difficult to circumvent the new privacy policy and prevent one’s data from being used for AI training. What we, internet users, can do for now in order to reduce the amount of our data available is share only information we’re comfortable with Google accessing or seeking alternative services. Until it is not illegal, Google will just continue updating its privacy policy in an at-first glance user-friendly manner – and there’s not much internet users can do about it.

 

Sources:

Étiquettes: , , , , ,