UiPath‘s George Roth discusses the transformative impact of large language models (LLMs) on artificial intelligence (AI).
Previously, AI has been inaccessible to the general public as a result of high training requirements and the expensive nature of resources. However, Roth believes that this is changing due to technological advances and the growing role of AI in business.
LLMs, such as the ever-growing ChatGPT, are allowing computers to understand human language and respond to a variety of complicated conversations and questions. Companies are using LLMs to gain valuable insights from massive language datasets, which can be fed back to the business.
However, Roth mentions that LLMs are too generic to be useful without customization. He argues that no-code model training tools can be used to give the LLM unique and specialized data, which it can use to make better predictions. Human subject matter experts can then complete the training by checking and correcting the model’s predictions. This creates a continuous feedback loop where humans are constantly training the AI to become better.
Model validation is also crucial to building trustworthy AI which UiPath has been enabling this through its active learning process, combining human and machine learning. Roth believes that democratizing AI will allow every employee to access it, unlocking countless new AI use cases across the enterprise.
Learn more about Roth’s takeaways here.
