Factiverse: Norwegian Startup Combats Disinformation with AI

The 2024 U.S. presidential election revealed the alarming speed of disinformation spreading online. This wave of false information influences public opinions on critical issues such as climate change, healthcare, and migration. The situation is worsened by generative AI, capable of creating deepfakes and fabricating facts within seconds.

Amid this challenge, Norwegian startup Factiverse has introduced an innovative approach. A participant in TechCrunch Disrupt Battlefield 200 in October 2024, Factiverse won the “Best Pitch” award in the “Security, Privacy, and Social Media” category. The company has developed a tool that helps businesses verify the accuracy of texts, audio, and video materials. The startup’s goal is to save time, reduce reputational risks, and minimize legal liabilities.

Since its founding in 2020, Factiverse has raised approximately $1.45 million in pre-seed funding. Despite being in its early stages, the startup already collaborates with leading media and financial institutions, including one of Norway’s largest banks.

Maria Ameli, CEO and co-founder, shared that their technology has already been used for real-time fact-checking during the U.S. presidential debates. Several media partners relied on these results for timely reporting.

A Unique Model to Tackle Disinformation

Factiverse stands out with its proprietary AI model trained exclusively on reliable, carefully curated data. According to Ameli, the startup deliberately avoids “poor-quality data” often used in training large language models (LLMs).

“Our model learns to think like a seasoned researcher,” Ameli emphasized.

Factiverse’s model can identify claims and analyze them in real-time, leveraging a wide range of sources: from search engines (Google, Bing) to AI-powered platforms (You.com) and academic publications. The technology doesn’t simply show the first search results but evaluates historical reliability and relevance to specific topics.

Facts Over Fabrications

Factiverse confidently competes with major AI models. According to Ameli, their technology surpasses GPT-4, Mistral 7B, and GPT-3 in identifying the factual accuracy of statements. Supporting 114 languages, their algorithms achieve an accuracy rate of approximately 80%.

“We are constantly improving these metrics while expanding globally to find new clients and partners,” Ameli noted.

The startup plans to raise seed funding in 2025 to scale operations and enter new markets. “We are seeking clients and investors who believe in trust and transparency,” the CEO added.