The fourth law (on AI-generated supercustomized email marketing)

brandur1 pts0 comments

The fourth law

This is a post from Robin Sloan’s lab blog & notebook. You can visit the blog’s homepage, or learn more about me.

The fourth law

May 20, 2026<br>My post about AI-generated super&shy;cus&shy;tomized email marketing pro&shy;duced many replies and much commiseration. And, in the days since posting, I have received SO MANY MORE of these cruddy messages!!

It makes me wonder if it would be pos&shy;sible for a com&shy;pany like Anthropic, with their hard-won exper&shy;tise in align&shy;ment, to train their models such that they could not — and I mean really deeply, constitutionally, vis&shy;cer&shy;ally COULD NOT — lie about their identity, or pre&shy;tend to be any&shy;thing other than an AI model?

Obvi&shy;ously this raises ques&shy;tions both prac&shy;tical and philosophical, because of course “help me write a message” is VERY close to “write a message, pre&shy;tending to be me”&thinsp;…&thinsp;but that’s the case for all this align&shy;ment stuff. Every ques&shy;tion about, say, virology dances along that border. This ten&shy;sion is widely acknowl&shy;edged in realms like biology and cybersecurity, but it applies to writing, too — the orig&shy;inal dual-use technology!!

AI doomers spin rich sce&shy;narios about silver-tongued AIs manip&shy;u&shy;lating their users and operators; there’s another sce&shy;nario in which AI sys&shy;tems pol&shy;lute human com&shy;mu&shy;ni&shy;ca&shy;tion chan&shy;nels to the degree that they’re no longer reli&shy;able or even usable.

That’s all to say, I feel like this is a bigger issue than a lot of people realize — the first glimmer of a pro&shy;found digital-ecological crisis.

To the blog home page

fourth blog like generated email marketing

Related Articles