Brendan Buckingham, Business Development Lead Data Fabric and AI at IBM, discusses the trends in the AI space and how they will impact businesses over the next 12 to 18 months.

He emphasizes the importance of using AI technology in a responsible and ethical way, and how it should be applied to real business needs. He also highlights the need for a variety of models to cater to different use cases, including general purpose models, domain-specific models, and fine-tuned models.

Other trends discussed include the importance of data privacy, the use of smaller models, the growth of AI assistance, and the need for AI governance and compliance.

Takeaways

• AI technology should be used in a responsible and ethical way, with trust being a key factor in its adoption.
• A variety of models, including general purpose, domain-specific, and fine-tuned models, are needed to cater to different use cases.
• Data privacy is crucial when training and using AI models, and data residency should be considered.
• Smaller models can be highly effective and cost-efficient compared to larger models.
• AI assistance, such as natural language interfaces, can enhance human performance and improve business processes.
• AI governance and compliance are essential for managing and monitoring AI models throughout their lifecycle.Hybrid deployment approaches, considering factors like data privacy and location, are becoming more common.
  Multimodal models and advancements towards AGI (Artificial General Intelligence) are areas to watch in the future.

Sound Bites

• “We need to put trust first, both with the people who are affected by this, but in the way we construct and use this going forward ethically.”
• “The future will involve using multiple types of models to cater for a wide variety of use cases.”
• “Bigger isn’t always better. Smaller models are highly tuned to specific tasks and are very, very capable.”

Chapters
00:00.  Introduction and Overview
01:21.  Putting Trust First and Ethical Use of AI
02:24.  Variety of Models for Different Use Cases
04:41.  Effectiveness of Smaller Models
06:15.  Domain-Specific Models for Enhanced AI
07:16.  Data Privacy and Residency Considerations
08:15.  Using Small Targeted Models for Efficiency
12:14.  Hybrid Deployment Approaches
15:27.  The Growth of AI Assistance
18:16.  AI Governance and Compliance
19:12.  Advancements in Multimodal Models and AGI
20:10.  Summary and Closing Remarks

We have launched a YouTube channel where you can also view our videos – just click here and please like and subscribe PTV Youtube