Norwegian startup Factiverse takes on disinformation with AI

Amid the disinformation surge following the 2024 US presidential election, Norwegian startup Factiverse is stepping up to combat the spread of false narratives.

Leveraging advanced AI, the company aims to equip businesses and media outlets with real-time fact-checking tools to counter misinformation across text, video, and audio formats.

Factiverse, which gained recognition at TechCrunch Disrupt Battlefield 200, was awarded best pitch in the Security, Privacy, and Social Networking category. CEO and co-founder Maria Amelie describes the startup’s mission as twofold: saving businesses time on research while minimizing reputational and legal risks in an increasingly misinformation-driven digital landscape.

Founded in 2020, Factiverse operates on a business-to-business model and has secured US$1.45 million in pre-seed funding. Early collaborations include partnerships with media outlets and major financial institutions, such as one of Norway’s largest banks. Notably, the company provided live fact-checking for the US presidential debates, a service utilized by several media organizations.

Factiverse’s AI model differentiates itself by focusing on high-quality, credible data sourced from fact-checkers and reliable publications worldwide. “We’re not an LLM,” Amelie said, referring to large language models like GPT-4. “We’ve built a different type of model based on information retrieval.”

The technology, developed in collaboration with co-founder and CTO Vinay Setty, an associate professor of machine learning at the University of Stavanger, uses machine learning and natural language processing to identify claims and verify their accuracy in real time. Factiverse’s approach scours a wide range of sources, from traditional search engines to academic databases, prioritizing credibility and relevance.

“We don’t just surface what comes up first on search engines,” explained Amelie. “We propose sources historically proven to be credible on specific topics, even analyzing who is quoted in an article.”

Factiverse claims its model outperforms GPT-4 and other leading AI systems in identifying fact-check-worthy claims and verifying their accuracy across 114 languages, with an 80% success rate. The startup aims to expand its reach globally while pursuing a seed funding round in 2025. 

Share this Post:

Accessibility Toolbar