fairsenseai.analysis.risk_assessment.analyze_text_for_risks

analyze_text_for_risks(text_input, top_k_risk=5, top_k_ai_rmf=1, embedding_model_name='all-MiniLM-L6-v2', progress=<gradio.helpers.Progress object>)[source]

Analyzes input text for AI-related risks and maps them to AI RMF guidelines using embedding-based similarity search.

Parameters:

  • text_input (str) – The user scenario text describing an AI project to be analyzed

  • top_k_risk (int, optional) – Number of similar risks to retrieve, by default 5

  • top_k_ai_rmf (int, optional) – Number of AI RMF matches per risk to retrieve, by default 1

  • progress (gr.Progress, optional) – Gradio progress bar object for tracking analysis progress, by default gr.Progress()

  • embedding_model_name (str, optional) – Name of the sentence transformer model to use as embedder, by default “all-MiniLM-L6-v2”

Returns:

A tuple containing:
highlighted_outputstr

HTML formatted string with highlighted risk entries.

temp_csv_pathstr

Path to the saved CSV file containing detailed analysis results. Returns an empty string if analysis fails.

Return type:

Tuple[str, str]

Examples

>>> scenario = "We're developing a facial recognition system for public spaces"
>>> highlighted, csv_path = analyze_text_for_risks(
...     scenario,
...     top_k_risk=3,
...     top_k_ai_rmf=2
... )
>>> print(f"Results saved to: {csv_path}")
Raises:

Exception – If there’s an error during the analysis process, returns error message and empty string as CSV path