We are living in an era of information overload, yet the ability to access facts has not become widespread. There exists an invisible barrier built from tools, methods, and experience between ordinary people and professional institutions. When opinions, propaganda, and data disguised as “facts” surge in like a tide, most people lack effective means to discern and defend themselves.

Diligence Check is designed to provide you with a convenient, efficient, accurate, and comprehensive information check experience. You can input textual content (such as news, social media statements, forum posts, content generated by AIGC, outputs from large models, etc.), or provide a webpage link. Diligence Check will automatically analyze the content to help you judge the reasonableness of the information, the reliability of the source, and the level of evidential support.

Service Scenarios:

• Reducing the risk of information misjudgment, helping you distinguish rumors and false information.
• Diligence Check aims to minimize the chances of making wrong decisions based on misinformation, empowering you to identify rumors and disinformation efficiently.
• Mitigating the spread of fabricated information, hollow conclusions, or misleading guidance introduced by model hallucinations, thereby improving the quality of scientific communication and the public opinion environment
• The tool helps prevent the propagation of erroneous or fabricated content, particularly those resulting from AI-generated “hallucinations,” thus fostering a healthier and more reliable media landscape.
• Providing a convenient information verification channel for media, researchers, and the public. Diligence Check serves as an accessible platform for fact-checking, catering to the needs of journalists, academics, and the general public alike.

How Diligence Check Works:

• Powerful Information Retrieval Engine:
  The platform integrates various search sources and webpage parsing tools to acquire a large amount of real-time information from across the internet, ensuring a broad and timely data foundation for verification tasks.

• Cross-Source Verification:
  It extracts key information that needs to be checked (including factual statements, data, citations, etc.), tracks the sources, and evaluates the authority of these sources. The system confirms whether the information exists and whether its usage is appropriate, classifying and quantifying multi-source information under a single topic, and presenting evidence clearly and conveniently.

• Source Authority Evaluation:
  Our system is designed to help the model understand human expectations for reliable sources during the fact-checking process. For instance, it strives to find primary, authoritative, and publicly accessible sources in professional fields. The model is trained to recognize differences in credibility across information channels, considering factors such as expertise, public trust, intent, and bias.

• Delivering Evidence, Not “Facts”:
  We do not let AI provide vague or generalized qualitative conclusions. Every fact-checking outcome is backed by multiple cited sources, original webpages, publication times, and other detailed information, ensuring transparency and traceability.

Are Fact-Checking Conclusions 100% Accurate?

Diligence Check relies on existing search engines and authoritative public information. For some very new or niche topics, there may be gaps in data coverage. We encourage users to question, supplement with more evidence and sources, and work together to build a stronger, more comprehensive information infrastructure and deliver better verification tools.

Future Plans—Broader Data Access, Stronger Reasoning Abilities:

• We will continue to update and expand the data sources for Diligence Check, and broaden the means of data acquisition, such as introducing tools like browser use to access private domain information.
• We plan to integrate these capabilities into our upcoming feature—In-Depth Research. In-Depth Research is an AI agent that autonomously completes multi-step, complex research tasks based on model reasoning abilities, helping you quickly accomplish deep and comprehensive research projects.