How can AI help investigative journalism? What are the challenges of introducing machine learning as one of the data analysis approaches used by journalists? Since 2017 ICIJ has been experimenting with machine learning as one of the data analysis methods used during its investigations.
In the Implant Files, machine learning was used to identify medical devices patients’ deaths that were misclassified when they were reported to the authorities in the United States. In Luanda Leaks, machine learning helped cluster records such as bank transfers and balance sheets.
However, the machine doesn’t always get things right. ICIJ has introduced several validation processes, including human validation and developed its own tool to help with fact-checking, making it possible to use the results.
There can be additional exchanges with the attendees through social networks and emails.
How will you deal with varying numbers of participants in your session?:The session welcomes any variation in the number of participants. Based on the number of people who are present, there would be different approaches towards interacting with the audience and facilitate the exchange of ideas on the central topic of the session.
Emilia is ICIJ’s data and research editor. She has taken part in cross-border projects such as the Panama Papers, FinCEN Files, among others. She coordinates partnerships in LatinAmerica.
Miguel is a web and data application developer in ICIJ’s Data Unit. He builds ETLs that process the data into more friendly shapes for ICIJ and its partners.