Are LinkedIn's AI insights a compliance headache for pharma marketers?

Artificial intelligence (AI) is impacting so much of our personal and professional lives, from cleaning up photos and summarising interview notes, to feeding cats (probably) and altering how we use social media apps.

However, when it comes to AI in pharma marketing, compliance concerns can make its use a little challenging.

For example, a funky new feature from LinkedIn called AI takeaways could be something of a headache for pharma marketers from an ABPI compliance perspective.

Here's how LinkedIn describes the feature:

“AI-powered takeaways on LinkedIn use generative AI to provide members with personalised insights and suggestions to help you find the right information and connect to additional resources on LinkedIn.”

In practice, this looks like a set of “buttons” beneath a post. Here's a screenshot:

When you then click on these buttons, they open a pane on the right of the desktop view to show AI-generated information and the opportunity to ask LinkedIn's AI further questions, like this:

At this point it's worth noting a few quirks of the feature. The documentation on LinkedIn's website is a little sparse, but we've got you covered with some key info.

Where do AI takeaways appear?

• This feature is only available to LinkedIn Premium users and those in certain geographies, including the UK
• AI takeaways don't appear on posts on the mobile app
• They also don't appear on promoted posts, so your paid campaigns are exempt
• They appear on both individuals' posts and company posts
• They don't appear on posts when you view them within a company's profile page, only in the newsfeed view (see, I said it was quirky!)
• They appear, from our observations, on upwards of 80% of posts in the newsfeed

Why do AI takeaways appear?

This is a little trickier, but from our research, these posts appear:

• When a company is tagged in a post
• When a clear topic, place, event, or initiative is mentioned in a post (such as “HCP audiences” in the example screenshot above)
• On a long-form post where LinkedIn can summarise what's in the post

If you spot any other situations in which these takeaways are triggered, let us know!

But here's the rub. If you're the creator/publisher of a post – you can't turn them off. And you can't impact what information they pull in.

What does this mean for pharma?

At first glance this looks like an interesting feature that proactively shares what could be quite helpful information, but it also opens up a number of pharma compliance considerations.

Funky final form

It might be argued that the way the AI buttons appear under the post, above the like/comment section, makes them feel part of the post – and so should be considered when the post is certified.

The snag is, as described above, these buttons are dynamic – so generating a final form screenshot for pharma review is simply not possible. There's also the nuance of them appearing/not appearing in different parts of the platform/app, so that's inconsistent as well.

AI-generated code breaches

In the majority of cases, clicking on an AI button will open up a right-hand pane that draws in content from an associated company website, usually with a hyperlink.

For pharma, this could result in users being signposted to website pages which have not been certified for them to view. A link to a global pharma website page being shown to a UK user is the obvious scenario (which we've seen).

Executive comms are impacted too

As mentioned above, AI takeaways appear on individuals' posts as well. So this is a consideration for any company spokespeople who are using LinkedIn to talk about their employer, as their posts also fall under the scrutiny of the ABPI code.

What's the risk?

As a brand new feature, there are no ABPI cases to draw a precedent from yet, but there are two schools of thought for medical signatories at either end of the ‘severity spectrum’…

1. These AI takeaways are entirely outside the control of the company, and any links, etc. are being presented by LinkedIn. There is no intent from the company to drive people to non-compliant content
2. These AI takeaways are generated as a result of the company choosing to publish content on LinkedIn, which makes the platform itself questionable as an appropriate tool for pharma marketers to use

What's a marketer to do?

Trying to game LinkedIn is a folly. Yes, you can avoid tagging a company in your post to reduce the chance of the AI picking it up and generating a link… but ultimately that's not going to be any guarantee of stopping them surfacing.

Or you might choose to only run paid ads from your account, which is how some local markets already operate (but obviously limits your whole LinkedIn platform strategy).

Another option, which is pragmatic if a little clunky, is to add a disclaimer at the bottom of your organic posts to say something like “Any AI-generated information related to this post is not controlled by [insert company name]”.

So, what's next?

We, and other organisations we know, have asked LinkedIn to consider allowing companies to turn off AI takeaways for the reasons above. We'll see how that goes!

We anticipate someone raising a code case, so we'll keep an eye on that as well, and update this post with any developments.

But in the meantime, we recommend engaging your compliance teams with all the facts, so they can have a proactive understanding, rather than receiving an unwanted surprise.