"Unlawful." That is the word Amnesty International used to describe what OpenAI, Google Gemini, and Midjourney are doing with your personal data online. The organization's full report lands like a gut punch, and the timing could not be more relevant for anyone who spends serious hours online, gamers very much included.
AI data scraping concerns grow
Get up to 80% off games only on GAMES.GG
Exclusive Discounts on Games
What the Amnesty report actually says
The report singles out some of the biggest names in generative AI and labels their data collection practices as a direct violation of privacy rights. The core issue is that these companies have been scraping enormous volumes of personal data from across the internet without meaningful consent from the people that data belongs to.
Here's the thing: this is not abstract. If you have ever posted a photo publicly, shared personal details in a forum, or left a comment somewhere online, that information may already be inside a training dataset somewhere. The report specifically highlights image generation as a particularly invasive corner of the problem. A photo you posted for friends could, in theory, resurface in some form through an AI model trained on it.
The privacy concern does not stop at images. AI chatbots now carry enormous amounts of personal information about users, and services like ChatGPT have already been integrating advertising into their platforms. The concern is straightforward: a system that knows a lot about you and is also trying to sell you things is a system with obvious potential for manipulation. That is not speculation, it is a structural incentive problem.
VPN search interest just hit a record nobody expected
The timing of the Amnesty report lines up with a significant shift in how people are thinking about online privacy. Global search interest in the term "VPN" hit its highest recorded point ever this past February, a 75% jump compared to February 2025 and a 334% increase against an average month back in 2010. Those are not incremental numbers.
VPNs can limit future data exposure but cannot undo scraping that has already happened. If your data is already in an AI training set, a VPN will not remove it.
The surge makes sense when you stack up the pressures. AI data scraping concerns are rising. More governments are pushing age verification laws that require users to upload official ID to access certain content. Regulatory environments in multiple regions are tightening around internet anonymity. Each of these individually would nudge people toward privacy tools. Together, they are producing record search numbers.
VPN usage hits record levels
Why gamers should care about this specifically
The gaming community lives online in ways that most people do not fully appreciate. Forum posts, Discord conversations, stream clips, profile photos, gameplay commentary, and years of social media activity all represent a significant personal data footprint. Generative AI systems trained on scraped internet data have potentially ingested a lot of that.
What most players miss is that the threat is not just about embarrassing images being replicated. It is about the broader picture: detailed personal profiles assembled from years of online activity, potentially used to target people with increasingly sophisticated advertising or, in worse scenarios, social engineering. The gaming community has dealt with targeted scams and phishing attempts for years. This adds a new layer to that problem.
For practical context on staying safer online while gaming, the gaming guides section covers a range of topics worth keeping in your back pocket. Checking game reviews before downloading anything from less familiar studios is also worth building into your habit, since malware delivery through fake game downloads remains one of the more common vectors for data theft.
Where this goes from here
The Amnesty International report is not a final verdict. It is a public pressure document, and its real function is to force a policy conversation that regulators in the EU and US have been moving toward slowly. The EU AI Act is already in phased implementation, and cases like this give enforcement bodies more concrete material to work with.
The companies named in the report have not issued substantial public responses at this stage. OpenAI, Google, and Midjourney have all previously defended their training data practices on grounds of public availability, though that argument is exactly what Amnesty's legal framing challenges.
For now, the clearest takeaway is that the gap between what AI companies are collecting and what users actually consented to is wide enough that one of the world's most recognized human rights organizations felt compelled to call it out by name. That is not a minor development, and the record VPN numbers suggest a lot of people already sensed something was wrong before the report made it official.
