My runner-up headline was: "Sex Sells, Even If It's
AI-Generated," and there were many other great options, but I went with the one above because I think it's an important milestone for understanding not just the impact AI-generated content is having
on people overall, but specifically, how it is influencing our perceptions of what is "real," or at the very least, what is human.
As I've written before, AI-generated content has long passed
the Turing Test, and traversed the Uncanny Valley, and the only thing separating our perception of what is AI vs. human-made -- watermarks aside -- is the context we perceive it in.
Omnicom
Media and Zefr's just-released "AI Slop or Not" report delves into a lot of that. Based on a recent
study of adult consumers in the U.S. and Canada, the report provides benchmarks for specific types of AI-generated content, including sexualized imagery, "brand boycotts," and other genres and formats
to readers of "3.0."
advertisement
advertisement
Even if they are just self-reported responses to a survey, I recommend that you at least peruse the report to see what people are thinking about AI-generated content and
how it may or may not be impacting brands, especially the adjacencies of their ads alongside it.
Much of it may be similar to consumer studies you've already seen, but I think we're witnessing
a progression of human adaptation to a world increasingly dominated by AI-generated -- or at the very least, AI-augmented -- stuff.
The reason I focused my headline, the chart above, and this
nut graph to what might otherwise be a throw-away finding from the report, is that I actually think it is the most substantive one, because it is one of the first benchmarks I've seen for how the
morass of AI-generated content is influencing the way people perceive human-created content. In particular, exacerbating skepticism for everything they now see.
Personally, I adopted that
perspective a long time ago, at least since the release of OpenAI's Sora 2, but I think it's an important benchmark to track over time, because I believe it will lead to greater skepticism over
everything people experience going forward.
Honestly, I don't think that's such a bad thing, because I think many people are too accepting of any form of content, especially in an age where
professionally-created genres -- news, information, entertainment and even advertising -- are increasingly being marginalized by the morass of user-generated and now AI-generated content and
curation.
In other words, I think the 32% of people who are now skeptical by default are the earliest adopters of the new media marketplace reality.
I may follow up and do some other
drill-downs on this report's findings, especially the section about ad adjacencies to AI-generated content genres. It's an important one.
But to wrap up today's post, I'd like to do something
I find myself increasingly doing these days: not use AI to generate content for me, but to create some with me. You know, collaborate.
The truth is that there are some things
AIs are better at than I will ever be. In particular, NotebookLM's ability to distill a 32-page report chock full of data points into a meaningful perspective already is quite remarkable.
And
some of the adaptations available in its "studio," including "mindmapping" and audio overview "podcasts" are things I
don't have the ability to do as well.
So in the spirit of man/machine collaboration, here's what NotebookLM has to say about Omnicom Media and Zefr's "AI Slop or Not."
