![](\"https:\/\/world.edu\/wp-content\/uploads\/2026\/09\/watermark-content.png\")
<\/p>\n<\/p>\n
Imagine a manager sends a thoughtful message to one of her hard-working employees after a particularly exhausting week in the office. The message communicates personal warmth, but when the recipient discovers it was generated by AI, it somehow feels less genuine.<\/p>\n
In the past, the recipient would only have discovered AI use if the sender disclosed it or through a lucky guess. Now, the European Union\u2019s new transparency rules<\/a> require companies to watermark content produced by generative AI systems in a machine-readable format. Systems already on the market have until December 2026 to comply.<\/p>\n OpenAI reportedly developed such a watermarking system<\/a> for ChatGPT years ago but held it back for fear of alienating users. Google now uses a watermarking system called SynthID<\/a> in its propriety AI models. Recently Anthropic followed suit<\/a>, announcing that content generated through its Claude models globally will contain an invisible watermark.<\/p>\n In other words, whether you reside in the EU or Canada or elsewhere, your AI use will no longer remain private or undetectable.<\/p>\n As researchers studying the impact of AI on learning, including the relational and emotional dimensions of that, this development concerns us deeply.<\/p>\n Recent research<\/a> demonstrates that although many people can\u2019t tell whether a personal message was written by AI or not, their perception of the sender changes the moment they learn it was AI-generated.<\/p>\n We tend to read the time and effort someone puts into their writing as signals that we matter to them. When that work is outsourced to a machine, it feels like some of that care is no longer there, even when the underlying emotions are genuine.<\/p>\n But the process of watermarking AI also feeds into an already prevalent culture of surveillance. This is especially evident in workplaces, where employees are more willing to accept a manager\u2019s use of AI for a routine scheduling note<\/a> than for a personal message that offers praise or feedback.<\/p>\n In the latter case, the manager\u2019s sincerity is more likely to be questioned. Watermarking makes this evaluative judgment available on demand. In a workplace where checking each other\u2019s messages for AI fingerprints becomes a habit, even casual communication can become marred with suspicion and doubt.<\/p>\n The AI detectors worldwide are about to get their biggest boost ever with new EU “watermarking” requirements that came into effect this month.<\/p>\n [image or embed]<\/a><\/p>\n \u2014 POLITICO Europe (@politico.eu<\/a>) August 14, 2026 at 5:29 PM<\/a><\/p><\/blockquote>\nA culture of surveillance<\/strong><\/h4>\n
\n