Priorities for Businesses Adopting Generative AI

The possibilities of generative AI excite both businesses and governments. The ease of interaction and the ability to turn your words into new content is exciting. So is the thought of engaging more interactively with the communities you serve. On the flip side, we don't know all the ways people will engage with the systems and how the system will respond to user prompts or what the impact will be on the users or the organizations deploying the systems. Generative AI models are evolving, and new approaches are emerging as we speak. The systems are brutal. We are building the road and the road signs as we drive. However, there are some non-negotiables for both public and private entities involved. Trustworthiness, safety, security, inclusion, and human-centric design. So it is upon the senior leadership and decision-makers in organizations to ensure the right mechanisms are in place when adapting generative AI systems. A starting point is to establish priorities. If you're a CEO or a newly minted Chief AI Officer, or an AI advisor in your organization, what should be your three top of mind issues to ensure trustworthy and responsible AI?

First priority is organizational readiness. Even though generative AI systems may support or even replace some of the tasks currently handled by humans in your organization, you'll still need AI-ready employees who understand the possibilities and limitations of the systems, who can conduct evaluations and validations independent from the vendors, who can go in these systems when they're deployed. Human agency and oversight are significant guardrails, not only because these are emerging technologies and need to be monitored, but also because generative AI systems are prone to inaccurate and biased results. So start with assessing your internal talent capabilities first and commit to investing in your human and organizational development in this process.

Second priority is due diligence. Generative AI might be a shiny object, or it can be transformational for your organization. However, it is hard to know the answer before conducting an assessment of your needs, compared to the possibilities of generative AI technology. Due diligence may come in different shapes and forms. You might assess what tasks and functions in your organization would benefit the most. Evaluate the value add from this technology to your current state or baseline. It might mean assessing your current technology or infrastructure, or the possible interactions and reactions of your clients with this new technology and product. Or, if you're acquiring a generative AI system from a vendor, due diligence means robust procurement due diligence. As an organization, public or private, you need to understand the risks and impacts and make informed decisions.

Third priority is governance. When you make a decision to develop or adapt the generative AI systems, you make a commitment to continuously monitor the system's performance and improve its functions. AI or generative AI systems are not static systems. So from the very beginning, you need to think of governance and accountability mechanisms, changes to existing roles and responsibilities, internal monitoring and control methods, documentation, and the communication flow for the tasks and decisions impacted by generative AI systems. You want to embed a responsible and transparent approach to your systems from the very start. You want to engage with the internal and external stakeholders, understand the risks, both positive and negative, and the impact of your proposed system on these groups. There are no one-size-fits-all solutions. You need to understand what works for your organization and those who you serve.