At InSight Crime, the integration of artificial intelligence (AI) into our workflow is governed by a commitment to maintaining the highest standards of journalistic integrity, accuracy, and security. This document outlines our public commitments regarding the use of AI tools in our reporting, data work, and multimedia production. 

Our AI standards are built upon three core principles:

  1. Informed and Transparent Use: InSight Crime is transparent about how and where we use artificial intelligence in our reporting process. Our journalists and investigators understand the tools they use, as well as their limitations, risks, and data handling processes. 
  1. Ethical and Responsible Conduct: Every story we publish is the product of human judgment, human reporting, and human accountability. AI is never considered an author, and we never present AI-generated content as the original work of our journalists. 
  1. Critical and Verified Analysis: Whatever an AI tool returns to us is raw material for our work, never the finished product. All writing, facts, figures, and claims are checked by human editors and reporters against reliable internal and external sources. AI may assist, but not replace, human verification in that process.

By adhering to these guidelines, InSight Crime ensures that technology serves as a tool for innovation while preserving the rigorous verification and critical analysis that define our organization.

What We Do Not Do

  • We do not present AI-generated images, audio, or video as authentic depictions of real events.
  • We do not consider AI an author, and we do not present AI-generated content as the original work of our journalists.
  • We do not enter our sources’ personal or identifying information into any AI tool.
  • We do not treat AI-generated outputs as verified fact or allow AI to replace human verification or specialized analysis.

Use cases

Data and Monitoring 

  • Scrape 400+ media outlets to monitor organized crime dynamics in the region. 
  • Processing and analyzing large datasets.
  • Coding. 
  • Collect, filter, and process news events to update datasets on criminal economies and criminal dynamics.
  • Transcribe interviews in multiple languages. 

Editorial

  • Proof read articles and social media posts.
  • Translate articles. 
  • Headline suggestions.
  • Contextual research assistance using an internal LLM fed with our daily regional monitoring and publications (background on actors, precedent events).
  • Assist in the drafting of social media posts.
  • Assist in fact-checking articles, profiles, and other editorial products.
  • Assist in the drafting of scripts and profiles using our contained, domain centered data.  

Multimedia 

  • Edit and create images. 
  • Edit and create illustrations. 
  • Edit and create basic graphics. 
  • Assisting in the drafting of video scripts 

Governance and Feedback

These standards are maintained by InSight Crime’s Monitoring and Data Analysis (MAD) unit, Managing Editor, the head of investigations and head of products, and reviewed on a regular basis to reflect changes in technology, risk, and best practice. Readers who have questions or concerns about our use of AI in a specific product are encouraged to reach out to info@insightcrime.org.