

The Top 5 AI And Data Trends 2025
Dec 12, 2024
In this engaging discussion, Bryan Harris, CTO of SAS, shares his expertise in AI and data science. He highlights five transformative trends for 2025, including the rise of synthetic data and the necessity of AI governance frameworks. The conversation delves into the challenges organizations face with regulation and energy costs while optimizing model training for sustainability. Bryan also explores the groundbreaking potential of quantum computing in AI, promising enhanced efficiency for industries like banking and healthcare.
AI Snips
Chapters
Transcript
Episode notes
Beyond LLMs
- Large language models (LLMs), while prominent, represent a small fraction of real-world AI needs.
- Other models, like fraud detection and document vision, are crucial for practical applications.
Synthetic Data Benefits
- Synthetic data offers key advantages: reduced data acquisition costs, enhanced privacy, and improved model accuracy.
- Some companies see 40-50% accuracy improvements by using synthetic data for model training.
Rare Event Modeling
- Rare events, like fraudulent transactions, pose a challenge for AI models due to limited training data.
- Synthetic data helps create more instances of these rare events, improving model robustness and accuracy.