Real-Time Fact-Check Layer for Your Feeds

Sources of truth.

👤 Source: Bill Kerr, founder and CEO of Athyna and Open Source CEO

❓ Problem

Misinformation is no longer a fringe problem, it is the headline. The World Economic Forum's Global Risks Report 2025 ranks misinformation and disinformation the single biggest global risk over the next two years, and it has now held the top spot for two years running. This is not climate, not conflict, not AI itself. It is the steady erosion of our ability to agree on what is true.

And people feel it. 58% of people now say they cannot tell what is real from what is fake online. We all have that one amazingly misinformed relative at the dinner table, and increasingly we are not sure we are not them.

Here is the part that makes this idea urgent. Just as the problem peaked, the referees walked off the pitch. In January 2025 Meta ended its third-party fact-checking programme in the US and switched to crowdsourced Community Notes, mirroring X. The trouble is crowdsourcing leaks badly: fewer than 9% of proposed notes on X ever get shown, and during the 2024 European elections only 11.7% of posts flagged by professional fact-checkers as misinformation ever got a note attached.

Meanwhile AI is pouring petrol on the fire. NewsGuard found that leading generative AI tools repeat false claims about 35% of the time, and they are now mass-producing fake images and video at zero marginal cost. So the volume of checkable nonsense has gone vertical, the people who used to check it have been laid off, and the job of sorting truth from fiction has been quietly offloaded onto you, mid-scroll, with no tools. Here's the idea.

✅ Solution

A browser plugin that sits over your social, news and content feeds and rates every checkable claim for factuality and bias in real time. A credit score for everything you read.

• It reads what you read. As you scroll, it underlines the checkable claims (not opinions, not jokes), so the signal stays separate from the noise.
• It shows the receipts. Each claim gets a confidence rating plus a left/right lean, with a hover-card that cross-references primary sources, surfaces the counter-evidence, and shows the date. No verdict handed down from on high, just the evidence laid out so you decide.
• It maps your media diet. Over time it flags your echo chamber, surfaces the other side of a story you would never have clicked, and turns the vague "am I being played here?" into a glanceable signal.
• It hands publishers a badge. Outlets whose copy checks out get a "verified" badge they can embed, closing the loop between readers who want to trust what they read and publishers who want a way to prove they earned it.

The wedge is narrow on purpose: one honest browser extension for people who are sick of being lied to. The platform underneath, a real-time claim-checking engine plus a continuously refreshed verification dataset, expands into publishers, platforms, and the AI models that suddenly need exactly this.

📊 Key Numbers

Market size

• The misinformation and disinformation detection slice of the AI-detector market is projected to grow from around $0.58 billion in 2025 to roughly $2.06 billion by 2030, a ~29% CAGR. The adjacent fake-image and deepfake detection market is on a steeper curve, heading toward ~$7.3 billion by 2030 at ~38% CAGR.
• But the real opportunity is bigger than "detection". This is the trust layer for a web where billions of people read news through feeds every single day. The TAM is closer to the entire media-credibility and brand-safety stack.
• Consumer proof points already exist. NewsGuard does an estimated ~$12M a year on a business model that is essentially "rate the news and license the data". Ground News has helped over 150,000 people break out of their echo chamber on a paid subscription.