AWS Solutions Architect Justin Lin: AI is evolving from generative assistants to intelligent agent AI systems, with model selection and data security being key.

ChainCatcher news, at the Silicon Valley 101 x RootData annual summit held in Silicon Valley, AWS Solutions Architect Justin Lin delivered a keynote speech, systematically elaborating on how enterprises can achieve a transformation path from proof of concept to large-scale implementation through AWS AI. Justin Lin pointed out that AI development is undergoing an evolution from generative assistants to intelligent agent AI systems, which will enable fully autonomous workflows and multi-agent collaboration in the future. He emphasized that model selection and evaluation are core aspects of enterprises applying AI. By 2028, most enterprises will deploy dozens of generative AI models, necessitating the establishment of a systematic testing framework and the design of architectures that support model switching. In terms of trustworthy AI, he cited data showing that 66% of corporate executives view data privacy and security as the primary risk, suggesting the construction of a security system from multiple dimensions such as identity management, encryption, and audit logs. AWS's three pillars of AI—innovation freedom, trustworthy AI, and value maximization—provide comprehensive technical support for enterprises. The speech also showcased a practical case from Crypto.com, where the accuracy of its AI assistant tasks improved from 60% to 94% by using Amazon Bedrock and SageMaker, providing second-level market sentiment analysis for over 100 million users, confirming the effectiveness of AWS AI in the cryptocurrency field.