The Alan Turing Institute talk about how they are using large-scale data analysis and cutting-edge AI development to provide real-time insight into the scope, prevalence and dynamics of harmful content online.  

This video shows how AI is used to gather information for online sources.  

Chapters 

• Introduction 

• The Online Safety Team 

• The Online Harms Observatory 

• Methodology for OHO  

• How Do We Train AI 

• Active Learning 

• How Does the Model Choose Entries?  

• Adversarial Generation of Data 

• Research Design  

• Test Set: Model Accuracy  

• Test Set: Model F1  

• Model results