Prisma helped uncover and mitigate inaccuracy caused by news articles in specific languages.
Purpose:
Osprey Flight Solutions has developed an AI system used to leverage its database of Aviation Specific historical data, the largest in the world, to identify any new anomalies or patterns that could indicate a risk to aviation anywhere in the world - usually days, weeks or even months before anything happens
To ensure the value delivered by their AI, and the quality of their predictions, the Osprey team wanted to assess the AI’s ability to correctly identify relevant aviation risks based on article news.
Key Insight:
Over the period of 4 months, the AI had a 61% accuracy rate on generating relevant incidents from articles. This dropped to 49% for news articles monitored in Urdu.
Conclusion:
A substantial ROI and business value was delivered to Osprey, enabling a more reliable and scalable solution that optimizes resources and improves decision-making globally in the aviation sector. It also provided them with a competitive edge through reliable and transparent AI deployment, reducing AI trustworthiness issues.