News app Artifact adds AI text-to-speech voices, including Snoop Dogg and Gwyneth Paltrow | TechCrunch

3 min read


Personalized news app Artifact, developed by Instagram’s founders, is adding another new AI-powered feature today. The company is launching an AI-powered text-to-speech feature in partnership with Speechify that will allow Artifact users to listen to news articles read aloud. The feature won’t just offer a robotic-sounding voice as some other text-to-speech engines provide, but will instead introduce a variety of natural-sounding voices that can be customized by selecting different accents and audio speeds. Two celebrity voices — Snoop Dogg and Gwyneth Paltrow — are also available.

To access the new feature, Artifact users will tap the play button now at the bottom bar of any article. They can then select the voice, accent, and speed to start listening. From a slider, you can adjust the speed as slow as 0.1x to as fast as 4.5x, though most will opt for something more reasonable.

The app will allow you to continue to browse the news while the article you’re listening to continues to play in the background. The company says this will make it easier to use Artifact to catch up on the news while you’re working out, commuting, or doing chores, for example.

All of the 30+ voices are free to use and there’s no plan to charge, Artifact notes. While they’re only available in the English language for the time being, users can choose from accents like the U.K., Australia, Nigeria, and South Africa. Though Paltrow and Snoop Dogg are the only official celeb voices, there are other fun voices to choose from. For instance, one voice dubbed “Mr. President,” sounds like Obama, while “Dwight” is meant to resemble Dwight Schrute from “The Office.”

The new voices are now one of many AI-powered additions that have arrived in the news app since its February 2023 public launch.

Last month, the company announced it would also use AI to rewrite headlines of clickbait articles. When a user flags a title as clickbait, the app calls on a GPT-4 model to rewrite the headline. These rewrites are marked with a star icon to indicate they’ve been changed from the original. And in April, the app began summarizing stories using AI, even letting people pick from fun styles like “explain like I’m five,” or even just a series of emojis, for a little whimsy.

The app’s recommendation systems are also powered by AI technologies to surface content personalized to the end user, based on factors — like their clicks, reading time, dwell time, shares, and more — instead of just broadly what’s popular across the user base.

Artifact is similar in some ways to other news apps on the market, including ByteDance’s Toutiao in China, Japan-based SmartNews, and News Break, another personalized news reader with Chinese roots. But in the U.S., it also faces steep competition from other sources where users get their news including the built-in news apps on smartphones provided by Apple and Google, and even social media sources, like TikTok.

With the addition of text-to-speech, Artifact will challenge other read-it-later apps like Pocket, Matter, and Instapaper, too.

Artifact is well-liked by those who have tried it, sporting a 4.7-star review across 4.3K ratings on the U.S. App Store, but it’s not yet blowing up to mainstream adoption, trailing the charts at No. 124 in the News category.

The new AI voices will roll out in an app update starting today.


Source link