In This Issue:

🤖 Meta announces its AI-powered creator tools, chatbots, and brand studio

🚫 Snap closes its enterprise AR development app

📹 YouTube will let more videos about sensitive topics have ads in them

⚠️ Yelp publishes “name and shame” list of businesses with paid reviews

🎮 Snapchat’s new game guiding agency staff through the pitching process

✂️ X slashes its Election Integrity Team and disables reporting of misinformation

✔️ X plans to add a ‘Verified Only’ reply option to boost paid subscriptions

🐀 Google is getting very, very bad. Just ask Futurama.

The Three Biggest Mistakes You Make on Google Ads [✨ Premium Exclusive]

Every Wednesday, exclusively in the Premium Newsletter and Premium Podcast, our Google ads correspondent Jyll Saskin Gales walks us through the latest platform changes. (Jyll runs the Inside Google Ads training program.)

This week: The three biggest mistakes marketers are making when setting up their Google Ads campaigns in 2023.

Meta: Is Its AI “Commitment” Real or a Flash in the Pan?

Meta held its annual conference today and reiterated its belief that the next big thing is NFTs… no, wait… the metaverse!… oh, hold on… no, it’s AI!

AI, of course, the topic which has captured the hearts and wallets of the techbros, and Meta has joined the club with a whole pool of announcements and we have them here for you.

Generative AI

First — and this probably won’t surprise anyone — more generative AI is coming to its apps.

Instagram will get something called “restyle” where you can type in something like “watercolour” or “magazine collage” and it will change your uploaded image to that style.

It will also get a backdrop feature that will let you upload a photo, then have the tool surround you with puppies, if you so desire.

If that sounds familiar, it’s because it’s similar to YouTube’s announcement a week or two ago — which will auto-generate a background for your Shorts. They call that “dream screen.”

Meta said it will indicate to viewers that AI was used, at least in part, though they’re still working on whether that’s a label or some other kind of indicator. TikTok already has some AI content labels in place, and YouTube is still figuring it out.

Meta is also planning stickers that you can create by typing out something specific — if you want a sticker of a unicorn birthday cake, just type that, and it’ll generate four options.

These stickers can be slapped on Facebook Stories, Instagram Stories, and within DMs in Messenger and WhatsApp.

Don’t go looking for it yet, though, this will start to roll out in the next month starting with a small group first.

Chatbots

Second, Meta has entered the the chatbot wars.

They plan to compete with OpenAI’s ChatGPT, Microsoft’s Bing chat, and Google’s Bard. But rather than being a web site to go to, Meta’s will be built into their messaging products in Messenger, Instagram, and Whatsapp.

It will be free to use and they have a partnership with Bing to be able to pull in real-time web data — something that ChatGPT charges for.

Also, Meta has licensed the likeness of a whole whack of celebrities like Charli D’Amelio, Kendall Jenner, MrBeast, Snoop Dogg, and Paris Hilton. There are 28 of these AI bots in total. Meta calls them — somewhat creepily — “embodiments.”

Others are themed-based like an AI chef and an AI travel agent.

AI Studio for Brands

Meta also is launching a platform that will let you build chatbots for your own brand and drop them into the message apps.

They call it AI Studio and expect brands will use it to connect with consumers and perhaps even handle customer service experiences.

It’s awfully similar to their one-time-obsession of bots on Messenger which were little more than simple data collection forms.

They spent months hyping the tools, the API, the customer experience — all to basically abandon it a couple of years later.

It’s hard to say what level of commitment Meta will have with this shiny-new AI polish on message bots. But you just have to ask the brands who invested heavily in Zuckerberg’s promise that the metaverse was going to be The Next Big Thing™ to get an educated guess on what the future might hold.

Then again, this technology has more of a track record than the message bots did — and one important feature of that track record: Monetization.

Other apps, like Snapchat, have proven that people will pay for add-ons that include AI chatbots.

(Side note: Snapchat is closing its enterprise AR tool. This was announced only back in March of this year. It let brands design augmented reality try-on features. Snap’s CEO told employees in an internal memo: “The advent of generative AI has made it easier for companies of all sizes to create try-on experiences for their customers and made it harder for us to differentiate our offering.”)

Meta’s AI Studio is, for now, just an announcement. Don’t expect to see it in the real world until next year.

Text to Image

One other addition to its new Message bots is users will be able to ask it to generate an image right from that chat. It’s similar to Midjourney in Discord, if you’ve used that. You type “/imagine” and then your prompt. This, too, will be free to use.

A company executive would not tell media what data the model had been trained on, but apparently it’s not trained on data uploaded by users to Instagram and Facebook.

At least for now. The Verge reported today that’s probably coming, though. The executive said “We see a long roadmap for us to tie in some of our own social integrations.”

YouTube to Permit More Ads Around Controversial Videos

There is other news today, so we turn now to YouTube, which has has long held a policy that certain video topics aren’t monetizable. Create a video about abortion, for instance, and it’s entirely likely that video will get monetization turned off — in other words, no ads can run alongside it.

The idea, of course, is that advertisers don’t want their brand names beside content like abortion, firearms, and sexual assault.

But now, that’s changing.

YouTube yesterday updated [see bottom of page] its rules around controversial issues — and now, ads can run alongside topics that used to be protected, as long as that content isn’t unusually graphic.

This is mostly a response to creators saying that YouTube applies those brush strokes too broadly — that a video discussing gun policy in the U.S. isn’t the same as one showing a 14 year old wildly shooting in the air while drunk. They’re both firearms videos, but they’re completely different in tone and should be treated differently.

But YouTube also clamped down on one particular topic — eating disorders — making it more restricted, and said it will no longer let ads run alongside videos that share information around binging or hoarding food or abusing laxatives.

This is, of course, a mostly net positive for marketers. By loosening up the guidelines on content available for advertising, while still keeping the really egregious ones out of the pool, that opens up inventory which generally lowers the advertising cost.

Inappropriate language

Violence

Adult content

Shocking content

Harmful or dangerous acts

Hateful & derogatory content

Recreational drugs and drug-related content

Firearms-related content

Controversial issues

Sensitive events

Enabling dishonest behavior

Inappropriate content for kids and families

Incendiary and demeaning

Tobacco-related content

Yelp’s Problematic “Hall of Shame” Web Site

Next time you ask your customers for a review in exchange for a gift card, you might end up on a new named-and-shamed list on Yelp.

The company has published a web page with the names of businesses it says it caught trying to “incentivize” customers for reviewing their restaurant or organization.

The page lists the name of the business, the location, then a link ominously titled “View the evidence here.”

One such piece of evidence was a screenshot of a Messenger conversation between a Korean BBQ restaurant and, presumably, an influencer.

How did Yelp get this private message? A note at the top says “A user sent us the below image of the offer,” suggesting the influencer narc’ed them out.

But wait a minute — nowhere in that message did the business owner say they’d comp their bill for a five-star review, or even a good review. They just said “A Yelp review based on your experience.”

Another example was from a tanning salon, which posted this text as an Instagram story:

Again, only asking for a review, not a “good” one.

To be fair, Yelp’s policy extends to asking for any kind of review, not just good ones.

This is something I’m not sure most marketers realize.

Certainly we all know about not asking customers for five-star reviews — if not the actual policy, at least basic common sense.

But how many realize they’re in violation of this broad policy of “Don’t ask anyone for a review at any time, for any reason”?

I’ll bet not many.

Google’s policy is probably closer to what most marketers would expect. Their Local Guides policy on deceptive content and behaviour notes that it’s against guidelines for a business to:

That, presumably, would allow for incentivizing reviews which are fair and accurate — a “genuine experience.”

Yelp says it has published notices of paid review attempts on nearly 5,000 businesses.

There are other naughty lists it’s publishing too — including a list of businesses where a bunch of positive reviews all came from the same IP address. They also put a pop-up calling it out on top of that business’s profile.

Snap Gamifies Agency Pitches

Snapchat has launched a quirky retro video game, which helps agency people design a Snapchat brief for their clients.

Well, calling it a video game is a little much — maybe a 16-bit interactive walkthrough is more accurate.

It gives you some insights on campaign building, some talking points about Snapchat’s offering, help with its ad products.

You can try it out at https://www.snapchatagencyadventure.com

There are actually four mini games.

Each highlights a stage in the pitch process, from the brief, to the team brainstorm, to building the deck, and, of course, the final presentation.

Snapchat says there are some easter eggs thrown in there — gamers, you know what I mean by that — a hidden augmented reality bonus game.

There are points involved, which means there’s a scoreboard, and Snapchat says as long as your agency is an official Snapchat partner, you can show up on the leaderboard.

In the end, you get a pretty decently fleshed out sample strategy for the app, with client goals and tactics.

That strategy is in all caps, using Comic Sans font.

Twitter Turns Off the Ability to Report Misinformation

Back to brand safety for a moment, and news today that X — formerly known as Twitter — has dramatically cut back on the team responsible for limiting misinformation and election fraud.

Media reports say about half the team which was left in that division was fired on Friday. And there were only 8 people left after Elon Musk’s previous layoffs. Now, there’s 4.

It wasn’t even a month ago when X said it would expand the team.

At the time, it said it would hire a civic integrity and elections lead, and hire people focused on information integrity. It’s not clear what the status of those hiring plans is now.

One of those laid off was a team lead in Europe. The Information reports that he had actually been suspended recently for clicking “Like” on a tweet critical of management at X — the app which says it’s devoted to free speech.

On Monday, the European Commission released a report finding that X had the highest amount of misinformation than any other major social media platform.

X Disables Misinformation Reporting

Also, an Australian organization has written X’s management demanding answers to why the social platform appears to have straight-up removed the ability to even report false information.

The reporting menu has things like spam, suicide, self-harm, violence, and so on or hate speech, among others.

But ‘false information’ is now no longer on that list of things users can flag for action.

Restrictions Coming to Post Replies

And, one brief note, X is working on a way to restrict who can reply to your posts. Right now, you can make it anyone, or just people you follow, or just people you mention.

An app researcher has found code showing that soon they will let you set your posts to only accept replies from accounts which are paying for the blue checkmark.

And finally…

There was a time when Google was good, remember?

It wasn’t even that long ago. You could type in something like “What is boneitis?” and it would give you back a list of links.

Up top, was usually a link to the wiki for the cartoon TV show Futurama, along with a list of videos from the show.

That made it immediately apparently that “boneitis” was a fictional condition created by the show writers.

But now, powered by the magic that is snippets… the sorcery that is AI… can we achieve even better results? Nay, dare I say, even more accurate? Perhaps a clear statement at the top saying “Boneitis is a fictional disease which was parodied in a TV cartoon?”

No, friends, we can not say that.

Because Google, in all its new-found knowledge and intellect, is now responding to some searches for “What is boneitis?” as if it were a real-world disease.

One Mastodon user today posted a screenshot of their search. Google’s reply at the very top of the screen — above all the links:

Wow, Google. Where did you get that information from?

Turns out, the text is a word-for-word lift from a Reddit comment nearly a decade old now:

What’s even more remarkable is that Google did not pick the top comment from that Reddit post. The top comment, as it turns out, is actually completely accurate, and reads:

When we searched Google for “What is boneitis?” we got a different, and somehow an even more inaccurate answer at the very top above everything else:

The photo beside this text, by the way, was a screenshot from the Futurama show.

By the way, we asked all the major AI chatbots the question ”What is bonetitis?” All of them — Bing, GPT 3, GPT 4, and Bard confirmed that the condition was fictional.

And who said the chatbots could never dethrone Google.

Some links in this newsletter may provide affiliate revenue to us.

This newsletter is powered by Beehiiv.