According to MIT, more than 140 big brands are buying ad space on dodgy AI-written sites

Covering culture, tech, luxury, travel, and media, twice a week


The new Mercedes-Maybach SUV’s absurdly spacious and luxurious back seat (Mercedes / Robb Report)

🔼 Bitcoin. It’s up 20% over the past 2 weeks »»

🔽 Georgia peaches. Farmers in the US state could lose up to 95% their harvest, thanks to wild weather »»

💬 “I didn’t think it would be this big.” Liquid Death’s 40-year-old founder describes how he turned ‘the dumbest name’ and a Facebook post into a US$700m water brand »»

🛫 20 top things to do in Oaxaca »»

👗 Kim Jones on his spectacular Dior show. Models appeared from below the floor »»

💎 The New Mercedes-Maybach SUV: “an opulent behemoth that’s as nimble as it is luxurious” »»


A Jesus with a neon halo in the Carlotta washroom. London’s newest Italian restaurant has been designed to evoke a sense of 80s glitz and glamour (Jérôme Galland / Dezeen)

Out: A-listers and big-budget audio narratives. In: audio talent that knows the medium down cold, and has more in common with talk radio or daytime TV hosts. In short, Hollywood’s leaving podcasting to podcasters »»

Victoria Beckham wore her most divisive summer footwear for the Jacquemus show »»

Taco Bell and Chipotle were named two of the most influential companies in the world »»

Zuck’s comeback is in full swing »»

Google’s head of search admitted that users aren’t pleased with the company’s search results »»

Reddit forum moderators call for an ‘affordable return’ for third-party apps in an attempt to break the standoff that’s caused thousands of forums (subreddits) to “go dark” or otherwise pause »»

More: as Reddit melts down, users are fleeing to lemmy, kbin, tildes and more »»

London commercial real estate update: HSBC is planning to swap its Canary Wharf HQ for BT's old digs in St Paul's »» More: A real estate reckoning is unfolding in real time »»

Also: Check out the decadent interiors of a new Italian restaurant in Marylebone called Carlotta. The whole thing is meant to evoke a 1980s Italian-American trattoria. Pro tip: Don’t miss the "midnight blue hideout" in the basement. The subterranean drinking den, complete with a 80s-style mirror-striped ceiling, will feature eclectic crockery and an open kitchen »»

A real life, hot-pink Barbie DreamHouse in Malibu is going to be avail on Airbnb »»

7 ways to save $$ on a trip you already booked —without canceling your vacation »»

These 5 killer Microsoft Edge productivity features are worth the switch from Chrome »»

Rihanna is the new face of LV »»

Wild: the social app IRL‘s users do not exist in real life. An internal investigation found that 95% of the app’s 20m users were “automated or from bots.” After raising more than US$200m in venture capital, IRL’s shutting down »»

Much of the internet already runs on Amazon’s cloud services, called AWS. The company’s long game? To create a single point of entry for firms to tap into a rapidly increasing number of custom generative AI tools. An “Amazon for AI,” if you will »» Related: OpenAI is also planning to create a marketplace for AI models »»


According to MIT, more than 140 big brands are paying for ad space on unreliable AI-written sites —and they probably don't even know it. What does that mean for the future of advertising?

The state of online advertising today?


For all its novelty, it can be easy to forget just how vital advertising is to the global economy. Really. Consultants note “a strong correlation between US advertising spend and real US GDP growth.” Worldwide, the sector is expected to be worth over US$1T in a couple of years.

So, what happens when one of the sector’s strongest pillars, online advertising, starts rotting?


Folks at MIT discovered that nearly 150 well known, large brands are spending money placing ads on sites filled with questionable content —much of it created by generative AI.

Stay with me as I break down why this is such a big deal.


Revenue is, of course, the lifeblood of any business. Businesses have options when it comes to increasing it, but the simplest, most straightforward one is to get more customers —and advertising, as everyone knows, is a time-tested way firms can reach new people, to do just that.

So, if this tool starts to lose its efficacy, well, every company has a problem.

But, as bad as that sounds, the issue goes deeper, still.


Around the world, including in the world’s richest economies, more people say that they trust businesses and brands than they do their own governments. By a huge margin.

In a survey of tens of thousands of people across 28 countries, businesses (brands) held “a staggering 54-point lead over governments when it comes to perceived competence —and 30 points ahead on ethics.”

Seen this way, this questionable content thing is actually a huge deal: if brands are trusted, then it makes logical sense that a brand's advertising output acts as a de facto as a seal of approval of the content it appears beside.

Brands advertising against demonstrably false, or risible, content is actually a bit of a crisis.


But, wait. Why and how exactly are brands spending hard earned money placing ads against questionable content?

The answer is “programmatic.”

“Programmatic advertising” is fancy marketing speak for when brands and agencies use algorithms to buy advertising space automatically and instantaneously. (This is as opposed to brands or agencies manually deciding to buy ad space on a particular site, or network of sites —in an old school “hey-everyone, let’s-sponsor-the-Yahoo!-News-homepage” way.)

Programmatic media buying lets brands/agencies purchase “impressions” (that’s just fancy online marketing speak for “views”) on sites, on apps, or really anywhere online instantaneously, in an attempt to get their ads in front of more eyeballs.

So far, so good, right? 

Well, programmatic’s automated nature can also mean that ads end up appearing in places brands don't want them.


This newsletter shared the shocking problem with programmatic ad buying last year.

Tldr: most programmatic ad buys end up sticking ads on sites that only bots view and click on. That’s horrible, and Uber famously got burned for this, wasting an astonishing US$100m on online ads that weren’t actually seen by people, which led to a flurry of lawsuits.

Meaning: This nonsense is not new. For years, programmatic ad buys have resulted in ad creatives appearing on sites that are actually meaningless fronts. The “clicks” they received? Meaningless signals of non intent, from non-human bots.


And it’s not just huge global firms like Uber who have been scammed.

Fate v Future reader Laura Lalonde, the founder of an innovative user generated content marketplace called Gridbank, shared a sobering “gut punch” of a story that anyone running paid ads will want to pay attention to:

She notes how “spam traffic has always been around.” But if you’re buying paid ads, “you're likely paying cash” for that traffic.

As she put it:

During one of my first ever marketing gigs, I dared investigate where (exactly) our ads were showing up.

4 meetings, 6 phone calls, the programmatic agency (one of the biggest in the world) FINALLY gave me the list of websites after calling me a ‘tough cookie’ ;)

Spoiler: the entirety of our ‘impressions’ were on bot junk sites.


The new news here is the scale at which people can pump out poor quality output/poor quality sites.

And it’s actually kind of scary. Here’s MIT again:

For example, one AI-written site, MedicalOutline.com, had articles that spread harmful health misinformation with headlines like “Can lemon cure skin allergy?” “What are 5 natural remedies for ADHD?” and “How can you prevent cancer naturally?” According to NewsGuard, advertisements from nine major brands, including the bank Citigroup, the automaker Subaru, and the wellness company GNC, were placed on the site. Those ads were served via Google.

In short, AI-written blogs are polluting the Internet, siphoning ad $$ away from better content —and potentially poisoning future versions of ChatGPT. (Or not.)


Fate v Future’s expertise goes beyond what you’re reading.

In addition to working hard to create the best newsletter on the Internet, we are a network of senior marketing experts, handling the trickiest tasks for brands both large and small.

As more and more spam traffic gets served to AI-written goop, we are seeing interest from clients in our new, non-technical SEO service: blog posts that are based on conversational or query-based long tail search terms/keywords— written with an especially human voice.

Pro tip: As a reader of this newsletter, you’ll get an insider rate on this service. Simply reply to this newsletter with “seo” to hear more.


Restoring trust in media and communications makes sense for a ton of reasons.

Remember the survey I mentioned above, the one that found that businesses are more trusted than governments?

It also found that the media is now the least trusted of the major four institutions (the other 3 are businesses, governments and NGOs). Trust in both traditional media and social media was down.

If, like me, you believe in the power of a strong brand —then you should also believe in the power of strong brand communications.

And that starts —as all good advertising does— with being truthful, and telling human stories people like and trust.

Bottom line: as all of the above plays out, look for the spread of AI to increase the value of HGC (human generated content), not decrease it.


Junk websites filled with AI-generated text are pulling in money from programmatic ads »»

“I have a gut-punching story for anyone running paid ads.” »»

“Brands are still more trusted than governments” »»

Ad industry growing at record pace, expected to hit US$1T by 2025 (from 2021) »»

Written by Jon Kallus. Any feedback? Simply reply. Like this? Share it!

Join the conversation

or to participate.