Today's media industry may not have a very positive view of AI. AI is a technology already being used to replace reporters with AI-generated copies, and other AI companies are scooping up journalists' jobs to satisfy chatbot data demands, without giving them back. . It brings traffic to publishers, just like search engines did in the past. But one startup, an AI newsreader called Particle from a former Twitter engineer, is changing the media industry by finding ways for AI to support publishers while helping consumers understand the news and dig deeper into stories. I believe that I can play a valuable role in this. ' business.
Backed by $4.4 million in seed funding from Kindred Ventures, Adverb Ventures, and other angels, Particle was founded last year by Sara Beykpour, Twitter's former senior director of product management. He worked on products such as Twitter Blue, Twitter Video, and Conversations. And he is the one who spearheaded the experimental app twttr. Her co-founder is Marcel Molina, a former senior engineer at Twitter and Tesla.
From a consumer perspective, the core idea behind Particle is to help readers better understand the news with the help of AI technology. Particle not only summarizes stories into important bullet points for easy catch-up, but also offers a variety of clever features that let you approach the news in different ways.
Image credit: Particle
But instead of simply siphoning off publishers' copyrighted material for its own use, Particle rewards publishers and drives traffic to news sites by prominently displaying and linking to the source directly below the AI summary. We aim to bring it back to .
First, Particle partners with certain publishers to host some of their content within the app via its API, including media outlets such as Reuters, AFP, and Fortune. These partners will be shown in a better position and their links will be highlighted in gold over other partners.
Image credit: Particle
Beta testing has already shown that the app's design and user interface are causing readers to click through to publishers' sites, but now the app is being released to the general public beyond news geeks. That could change the situation. In time, the company plans to introduce other ways to work with media in addition to sending referral traffic. The team is also in discussions with publishers about providing users with access to paywalled content in a way that makes sense for all parties.
“Forging deep partnerships and collaborations is one of the things we're really interested in,” Beykpour says.
To aid your traffic referral efforts, the article section of the app includes large tap targets, making it easy for readers to click through to the publisher's site. Additionally, Particle's byline includes the journalist's face, and readers can follow a link to the publisher's profile to read more content or follow them.
Using the app's built-in AI tools, news consumers can toggle between different modes such as “Explain it Like a 5-Year-Old” to deliver simplified versions of complex stories or “just the facts” summaries. You can get the version (or 5W). who, what, when, where, and why). Get news summarized in another language other than English, listen to audio summaries of articles and customized article selections on the go. Particle can also pull important quotes from stories and other helpful links.
Image credit: Particle
But two of the more interesting features include how Particle leverages AI to present news from different angles, and how you can engage more deeply with the story at hand by asking questions.
At Particle, one of our tools, called “Opposite Sides,” aims to burst users' filter bubbles by presenting different perspectives on the same story. This model has been tried before by other news apps, including startups Brief and SmartNews. Unlike previous efforts, Particle includes a story spectrum that shows how news is covered on both “red” and “blue”-leaning sites, allowing you to determine the position of the news. There are bubbles placed to indicate how far to the left or right and how large the news is. Reports may come from either side. AI also summarizes both sides and allows news consumers to reach their own opinion on the issue.
Image credit: Particle
However, the app's killer feature is its AI chatbot that lets you ask questions about your story and get instant answers. The app includes suggested questions and questions from others. For example, if you're reading about President Trump's immigration policy plans, you could ask the chatbot a question like, “What are the potential legal challenges to Trump's deportation plan?” or “What are the potential costs of mass deportation?” among others. Particle then uses AI technology to find those answers and fact-check them for accuracy.
“The chat feature uses OpenAI and our own pre-processing and post-processing,” Beykpour explained in an interview with TechCrunch. “If you want to use your content and find additional information on the web, search the web a little bit and generate the answers,” she says. After the answers are generated, Particle uses an AI to It includes an additional step where you have to go find support materials that match.
Overall, the app includes technologies such as OpenAI's GPT-4o and GPT-4o mini, Anthropic, and Cohere, including Google's non-LLM-based traditional AI technologies.
“We have a processing pipeline that takes the relevant content, summarizes it into bullet points, headings, subheadings, and does all the extraction,” she continues. “And…we pull out quotes and links and all sorts of relevant information. [the story]. And we use a proprietary algorithm to rank them, so the most important or relevant links are the ones that appear first. Or we'll show you the links that we think are the most important or relevant quotes first. ”
The company claims its technology can reduce AI accuracy problems that would normally occur once in 100 times to as low as 1 in 10,000 times.
As Particle grows, it also plans to use human editors to better manage AI content and homepage management, she said.
The app is currently free to download on iOS and works on iPhone and iPad.