• Meta ran an ad touting Muse could let you see who unfollowed you on Instagram.
• Instagram has long prevented its users from doing that. Meta has since removed the ad.
• This shows a complicated tension for Meta between the goals of AI agents and its social media platforms.

Muse, the AI agent from Meta, can do many wonderful and useful things. It can help organize your inbox, book travel plans for you, and buy you lunch.

But one thing that Meta has decided it shouldn't brag about is that it can track who unfollowed you on Instagram.

About a week ago, Rachel Karten, who writes the excellent marketing newsletter Link in Bio, noticed a strange ad on Instagram. The ad encouraged people to download Muse, with an image of the fuzzy mascot holding up red and green Tinder-esque signs, suggesting you could "See who unfollowed you each week."

Finding out who unfollowed you on Instagram is something people have wanted to know since the app first launched. But Instagram — for good reasons — has avoided making this a feature on its own app.

I reached out to Meta to ask about this seeming clash, and a Meta spokesperson confirmed to Business Insider that the ad had been removed and is no longer active in the Meta ad library.

Instagram doesn't want you to know your unfollows

Here's the tension: Instagram knows that people want this feature, but it also knows it would be bad for the platform's vibes.

Instagram head Adam Mosseri regularly does AMAs on his personal account, and he's answered questions about why there's no "who unfollowed me" feature. His reasoning (gone from his page but have been aggregated in other videos) is that they want to make people feel comfortable unfollowing people they're no longer interested in.

To most people, it is obvious that this feature could hurt people's feelings, especially younger users (let's assume that Instagram does care about teen users' feelings here, although I understand if you have some hesitations on that). Feeling bad while using the platform is a bad user experience!

It's also ultimately bad for Instagram. If you're reluctant to unfollow accounts you don't like, then you'll end up with a feed that's less compelling, and you'll end up opening Instagram less often or spending less time on it. That's bad for Instagram's daily active user metrics!

Yes, Muse can actually do this

There are, and have been, third-party services that can track who unfollows you. These sometimes require access to your login info in ways I definitely wouldn't recommend, and can track over time who has unfollowed you. So it's not like Muse is doing something totally novel or unheard of here.

I asked Muse to do this for me, and it said it could — with a caveat: it will start tracking my follow lists to see who unfollows me going forward, but it can't retroactively check who unfollowed me in the past (that's mostly the same as the other trackers). And honestly? I don't want to know! No thanks! Yikes!

Is this just one stupid ad? Yes. Was it set up and run as an ad by someone at Meta far removed from Instagram's philosophy on unfollows? Probably. Is it the most shocking news in the world if Meta runs a single ad and then takes it down? No.

But this interests me so much because it gets at a bigger and thornier issue. AI agents like Muse are what Meta clearly sees as the future of the company and how people use its services. But AI agents' abilities and use cases may be at odds with the nature of social media.

In the simplest way, AI agents are supposed to help you be efficient and get things done. Social media is supposed to help you waste your time.

These are two vastly different objectives — and it's going to be interesting to see how these two new technological pillars of Meta clash or complement each other.

Already, we have one Meta product (Muse) that is more than happy to do something that another Meta product (Instagram) has long discouraged.

There are plenty of other features that people want to do on social media, but Instagram and Facebook don't allow because it's bad for the overall health of the platforms. Suddenly, AI agents like Muse can do these long-forbidden things. The question is: how far is Meta willing to cannibalize the health of its own social platforms for the sake of growing its AI?

• Telehealth company Ro is launching an AI agent trained on customer data.
• The agent can recommend prescription changes, schedule deliveries, or write memos to a doctor.
• Eventually, Ro hopes to make it a "quarterback" for care. Will it improve outcomes?

Wouldn't it be great if we all had a little, digital minion handling all the small back-and forth chores of our medical care?

On Thursday, Ro is officially announcing the launch of its new AI health agent, designed to do more than a health inquiry on ChatGPT. It's the latest in a long lineup of healthcare agents — from insurers, telehealth providers, electronic health record systems, and regular chatbots — promising to cut down patient wait times, increase access to care, and improve personalized health recommendations with AI.

Ro CEO and co-founder Zach Reitano gave Business Insider an exclusive preview of Ro's new agent, called Robin, showing us how the tool could be used to perform routine tasks like managing side effects, meal planning, or making medication changes.

A founder's mission: disrupt healthcare delivery with AI agents

Reitano said this isn't just about making life more convenient for Ro's current subscribers, though that is the starting point. His long-term goal is to identify pain points where Ro customers need to seek out specialists for common conditions or prescriptions, then start offering more and more of that care in-house, becoming your go-to "quarterback" for the entire American healthcare system.

"I think it's very unlikely to me that one provider or one system handles 100% of someone's care, but I do think it is quite likely that there is one place in someone's life that serves as that quarterback and helps them navigate the rest of the healthcare system," he said. "We hope to be able to just continue to increase those concentric circles, where instead of saying you might be better suited outside of the Ro ecosystem, we would want you to be better suited in it."

When I asked him if his "ultimate goal" is to become a full-service virtual clinic, he said yeah, that's the hope.

For now, Ro's AI agent uses information that's already stored in the app, including notes from doctors, patient feedback, prescriptions, and Apple Health data to try to personalize recommendations for medication changes, dosing tweaks, or habits that might improve side effects.

It's meant to be more instinctively knowledgeable about your health than ChatGPT — requiring less prompting or uploading health data, since it already has information about patients on file. It can suggest a change in medication, and then go ahead and write a first draft of a note to your doctor, if you want it to.

"The underlying premise is: If I have your underlying health information, I can provide more informed and better things, better recommendations," telehealth researcher Dr. Ateev Mehrotra, chair of the health policy department at Brown University, said.

Only issue is, we don't yet know if that's true.

"People are giving up a lot of data for this, and so you really have to say that it truly is going to give you something," Mehrota added. "It sounds cool in theory, I'm not saying it's a bad idea. I would love some evidence before millions of Americans start doing this."

Personally, Mehrotra says he'd be much more likely to try out one of the AI agents embedded into an official electronic health record system, which has access to everything he's told his regular doctors and specialists within a specific hospital system and which — critically — has to abide by strict privacy rules. (Business Insider did not see Ro's privacy policy for Robin.)

"My health system doesn't have the weird financial incentives that we're worried about with Ro," he said. "More data, less financial incentives, I might use that."

Reitano says ultimately, it's not in his company's interests to go after short-term profits at the expense of a good patient experience.

"We fight for patients, and we serve you, and if we don't, then we go out of business," he said.