I’ve been getting AI generated spam for well over a year. It was immediately clear to me when it started. My spam emails became slightly more personalized than regular spam. They were all short: usually 2-4 sentences. The topics seemed to come in waves, all vaguely relevant to the owner of a small business or someone in marketing: there was the virtual assistant spam, the “do you want to sell your business” spam, and – my favorite – the AI generated spam selling AI generated spam tools. Most importantly: they were no less annoying to me than regular spam; unwanted and unsolicited interruptions in my day requiring me to manually mark them all as such.
Then, last week, something new happened. I got a very poorly targeted email from a life science company:
The notion that someone in life science marketing would want to buy genomes and metabolic pathways is ridiculous, but the real revelation was that the AI generated spam has penetrated into the life science market! This made me wonder if it’s changed people’s opinions about spam: after all, the whole point of AI generated spam is to replicate the more effective elements of one-to-one cold emailing. Perhaps improved personalization and relevance actually do make people more receptive to it.
Survey time!
The only way to answer the question is to ask. We posted a simple poll to the LabRats subreddit asking if they get AI-generated spam from scientific suppliers. I don’t think the result should be considered surprising:
A little over half the respondents report getting AI generated spam from scientific suppliers, and of those people almost all of them dislike it as much as regular spam.
What should we learn from this?
AI isn’t a magic bullet. It just makes bulk unsolicited emails a lot easier. Rented lists and low-cost bulk email service providers did too, and a lot of companies used them until deliverability plummeted and marketers realized that the costs to their brand’s reputation weren’t worth it.
Cold emails can be highly effective when executed correctly, with genuine, meaningful personalization and hyper-targeted sales pitches. It’s probable that AI sales tools will get to the point where they can do that, but the current iterations of generic AI sales tools just aren’t there. Like the bulk spam before it, we expect that AI spam will be increasingly, and preemptively, relegated to spam folders as mail servers slowly but surely learn that no one wants it.
Nothing is better for customer retention than great products. As marketers, however, the quality of the product is at least somewhat out of our control. The easiest tactic that the marketer has to improve retention is, ironically, one of the ones that can most easily turn customers off: email.