Andrew Hewitt, Principal Analyst
It’s a classic “be careful what you wish for” scenario. Executives are encouraging overworked, burnt-out employees to embrace AI as a solution for productivity challenges. But they may wind up with more than they bargained for as workers bring their own AI solutions into the enterprise, whether sanctioned or not. Welcome to the era of bring-your-own AI, or BYOAI.
In this episode, Principal Analyst Andrew Hewitt joins to discuss the benefits and risks of BYOAI. The discussion begins with Hewitt defining BYOAI (employees using any form of external AI service to accomplish company-related business, regardless of whether it’s sanctioned by the business) and differentiating it from shadow AI (which is BYOAI without the business knowing about it). Hewitt emphasizes that both cases include not only generative AI solutions but any software with embedded AI, which even the employee may not be aware of.
The conversation highlights the many nuanced concerns AI is creating in the enterprise. On one side, as many as a third of workers are concerned that their roles will be replaced by AI or some form of automation driven by business leaders. But on the other side, executives are increasingly concerned about the risks brought about by broader use of unsanctioned AI in the enterprise. At the extremes, some companies have issued outright bans of generative AI tools like ChatGPT, while others have bought licenses for every employee to encourage its use before weighing all the consequences.
Hewitt unpacks a number of the risks associated with BYOAI, including traditional security risks (e.g., data loss, IP leakage) that come with the use of consumer AI, legal risks related to bias and discrimination though the use of AI, and the risk of using third-party applications that have generative AI embedded without the organization’s knowledge.
So what’s the answer? Hewitt suggests starting with a BYOAI policy that guides usage but also evolves and is more empathetic toward employees than the restrictive BYOD policies were. A BYOAI policy should also clearly outline the approved use cases and “tell employees what’s in it for them — what types of things can they expect to be able to do with this generative AI service?” He also says that tech leaders should make sure they involve other teams, such as security and data governance, in the development of those policies and include employee feedback.
The episode concludes with a discussion of the impact of AI’s growth on the chip market and the move toward AI-enabled chips in PCs to support the growing demand for generative AI capabilities. To learn more about the impact of AI in the enterprise, be sure to check out Forrester’s 2024 predictions.