The Truth About the Death of Artifact

Credit: DALL-E

The fact that Artifact, the AI-powered news aggregation app launched to much fanfare last year by Instagram co-founder Kevin Systrom, announced it was shutting down about a year after launch, could be taken as a strike against the viability of using AI in news — or more specifically, news distribution. But that would be a misread of what really happened.

Artifact made a name for itself with AI-powered features like bullet-point summaries of articles created and cloned celebrity voices that could read you the news. The Goop website doesn’t let you listen to articles in Gwyneth Paltrow’s voice, but Artifact does.

Artifact offered novel, AI-powered features, like reading articles in cloned celebrity voices.

By most indicators, Artifact was well-liked by its users. Not only does it enjoy a 4.7 rating in the App Store, but in the aftermath of the shutdown announcement, several users on X logged their disappointment, noting how it quickly became a favorite on their phones. That’s a self-selecting group, to be sure, but if a product has deep flaws, usually you see at least a few folks pissing on the grave. That didn’t happen with Artifact.

If Artifact did anything wrong it was having expectations in line with broader AI adoption trends, which all point to hockey-stick growth in both business and cultural landscapes (look no further than this newsletter for evidence). The app was stymied not by a lack of vision or innovation, but by the media industry itself, which has never been kind to news aggregators as standalone businesses. Even the ones that succeed typically end up being niche products, and that’s frankly not good enough for the likes of Systrom, who said as much in his Medium post.

Subscribe now

That’s not to say Artifact is a failure. I’d argue that, more than any other service, Artifact showed tangible ways to apply generative AI to news consumption. The features it launched worked well, and it even managed to hit the sweet spot with notifications — when to send them and what stories to highlight — an impressive achievement on its own.

In short, Artifact was a great testing ground for AI-powered news, and some larger newsrooms I’ve talked to have thought about replicating some of its features on their apps and websites. After all, good features are good features: generative summaries have clear utility, news sites’ equivalent of “jump to recipe.”

Ultimately, though, Artifact shows the futility of starting with the solution instead of the problem. AI is a great tool, but it’s not a goal for either readers or publishers. At least now we know how much runway you get by leaning on it as your sole differentiator: about a year.

The Media Copilot is hosting a meetup on Feb. 1 in New York City! Join us for an evening of conversation over drinks with journalists, PR professionals, and executives about AI’s role in media and the news. RSVP to reserve your spot.


Ready to start using AI like a pro?


Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.