ChatGPT and generative AI have emerged as some of the most disruptive technologies in recent years. While end users find these technologies exciting, it is important to understand how companies view them from an enterprise level. This article explores the real business use cases and risks associated with chatGPT and generative AI, and how companies are handling and adjusting to these technologies.
Enterprises are facing the challenge of adopting new and untested technologies like chatGPT and generative AI. The landscape of technology has drastically changed with the shift towards cloud-based and interconnected systems. For example, Microsoft, which owns open AI, the creator of chatGPT, is already using this technology in its cloud-based version of Microsoft Office. This raises questions about the integration of chatGPT into Microsoft’s Office products.
Companies are scrambling to understand the implications of chatGPT and generative AI. Upper management, including CSOs, architects, and third-party advisory companies, are engaging in conversations to gauge their company’s position compared to others. They are focusing on use cases, risks, and concerns related to these technologies.
One of the risks associated with AI is hallucination. AI is designed to provide answers, but sometimes it may come up with incorrect or misleading responses. This can be a result of the data used to train the AI models. The black box risk refers to the challenge of understanding where AI is getting its information from. To address these risks, companies are working on explainability toolkits that provide context and references to the answers generated by AI.
Another concern is the use of copyrighted information by AI. Companies need to ensure that the data used by AI models is properly licensed and does not infringe on copyright. Permission-efficient data sources are being explored to address this issue.
Despite these risks, there are numerous business use cases for generative AI. In the software engineering domain, chatGPT can assist with code generation, code analysis, code search, and documentation. In the natural language domain, it can be used for text generation, Q&A, summarization, and language translation.
Financial services can benefit from generative AI in pattern recognition, trend analysis, and code generation. Companies like GitHub and AWS have already introduced AI tools for code development and generation.
Companies are implementing AI tools by following different approaches. Augmented generation involves using privately available data within the organization and augmenting it with publicly available data. Fine-tuning allows companies to customize the data used to train AI models. Private only modeling ensures that the AI models are trained only on private data, but it can be challenging and expensive.
Despite the risks and challenges, enterprises are responding quickly to integrate chatGPT and generative AI. This article highlights the concerns and progress in this space. Watch the accompanying video for further insights on the concerns over chatGPT and AI.
