Frequently Asked Questions
Answers to common questions about how SussMyAI works and what you can expect.
Your question is sent to multiple large language models (LLMs) through an Application Programming Interface (API). Their responses are then passed to another LLM through an API and compared. An overall “Sus Score" is assigned that measures how well the answers agree or conflict. We also check for the latest Google news updates (through SerpAPI) and look for any key differences between these and the AI model responses.
The question you enter is sent to GPT 4.1 Nano (served by Open AI), Claude 3.5 Haiku (served by Anthropic), Meta Llama 4 Maverick (served by Together.ai), and Gemini 2.5 Flash (served by Google) for their responses. These responses are then sent to GPT 4o (served by Open AI) for comparison and scoring.
No. While we highlight potential inconsistencies, you should always double-check our Sus Score and AI answers manually. SussMyAI is meant to support—not replace—your critical thinking.
“Sus” stands for “suspicious.” A High Sus score means the models disagree on key points, which suggests the AI responses may be inaccurate or hallucinated. A Mid Sus score means at least one of the models disagrees on a key point, so further investigation is recommended. A Low Suss score means all models are in relative agreement on key points. "Suss" means to investigate something so as to gain more knowledge.
We do not store any user questions or prompts long term. Only minimal logging of IP addresses and timestamps is done for analytics and security, as described in our Privacy Policy. However, please note that the LLM providers we use (and SerpAPI) retain the right to store user submitted content for their own use including training their LLM models. Also, if your submitted content is flagged for abuse or violation of our terms of service it may be logged. Please review our Terms and Conditions for more details.