seer. has had more than 700k views highlighting the dangers of AI Hallucinations in mainstream Internet Search. We believe that a safe and ethical integration of LLMs into Internet Search/Social Media should at least include factual accuracy (as does the AI Verify Foundation) if not transparency of the LLM’s sources and reasoning too.

As the LLM app creator of www.facticity.ai which is the most demonstrably accurate (review the datasets below), accessible, timely, explainable, and meta-analytic (and soon multilingual), benchmarking/testing is important to prove our unique advantages.

Dataset Accuracy

Fact Checking Tool Accuracy (%)
Facticity.AI
92%
Perplexity.AI
88%
Bing Copilot with ChatGPT-4
78%
Originality.AI
72%

Video

Benchmark Dataset

Multilingual Dataset v2

Multilingual Dataset

Updated and Expanded
Originality Dataset Benchmark

Benchmarked as of 11 June 2024 based on the assumption of using true and false as the only two categories with inconclusives as false Dataset available on https://aiseer.co/facticity/

*Note: Some labels from the Originality dataset were altered to correct for outdated information and faulty label extraction (e.g. “Chat GPT can access up-to-date information” was false but is now true with GPT-4)

It’s more than searching

Facticity verifies claims and debunks misinformation instantly with credible sources. Stay informed with trustworthy information.

Versatile Fact-Checking

Upload text, audio, or video content directly to our platform for a comprehensive fact-checking experience.

Our Clients

Subscribe to our newsletter