The AI opt-out models Meta, Musk's X, and the UK gov are proposing are simply not a good enough way for us to protect ourselves from data scraping

Nobody likes the idea of what major social media companies do with their information but as trends like "Goodbye Meta AI" suggest, people are even more worried about what AI scrapers, specifically, are doing with their data. Proposed changes by the UK government and Elon Musk's X, like Meta before them, could end up being pretty tedious to opt out of, if not downright obfuscated.

Starting with X, as spotted by Tech Crunch, a recent change made to the privacy policy of the social media site says it may share your data with third parties. If you don't opt out of this data sharing, it can be used to train AI models, "whether generative or otherwise". You can opt out of this by going into 'Settings', then 'Data Sharing and Personalization', and then turning off data sharing.

This is turned on by default and you are not warned of such upon creating an account. However, making an account does entitle the social media site to harvest data used on it, so it doesn't appear to be under any obligation to do so. Not only does X have its own AI model and chatbot named Grok, but your data could, and likely has been, used to train AI models from other sites.

In a very similar story, as reported by the Financial Times, the UK government is currently consulting on a proposal that would allow companies to train AI models on data scraped from websites unless its users choose to opt out.

This is frankly not a good enough way to allow consumers to be fully educated in how and why their data is used. Giving users the ability to opt out of AI scraping in some fairly obscure part of an app's settings won't give the majority of users enough information to know how their data is used and that they can even opt out in the first place.

An opt-in model would work much better here, where users can choose to allow their data to be scraped by AI if they so choose. However, it's hard to believe enough users would do so to satiate the data desires of current AI models and their owners.

If site owners are unhappy with an opt-in model, that implies owners are aware many would not choose to have their data scraped, and this touches on part of the problem with opt-out policies. It feels like an appeasement to those in the know but not a good enough tool to alert the average consumer about their own data rights.

AI scraping specifically, is a very new thing for many users and sites, and it feels like many of the current models are built out of a sort of Wild West, frontier approach which was there at the very start of generative artificial intelligence. AI companies have been bypassing copyright and acting in ethically ambiguous ways to get data for some time, so consumers need to be more proactive about their privacy than ever right now.