Playing by the rules of AI

The Fast Company Impact Council is a private membership community of influential leaders, experts, executives, and entrepreneurs who share their insights with our audience. Members pay annual membership dues for access to peer learning and thought leadership opportunities, events and more.

The future of AI is bright, yet its continuous evolution and an uncertain regulatory environment cloud its reality for many businesses. The new year has begun with a more relaxed stance on AI policy and DeepSeek’s supposed “better mousetrap,” giving CIOs pause on which direction to go.  

But these are good reminders for organizations to proactively establish best practices for AI implementation. Doing so will ensure future compliance while effectively leveraging AI for transformation and growth.

How regulation may impact innovation

In the U.S., California legislation has taken a first stab at defining guardrails domestically, following the footsteps of the EU Artificial Intelligence Act. However, these early regulation attempts will take time to enact and be vetted for successes, failures, and inevitable adjustments.

We expect AI to be governed differently according to three broad categories:

1. AI creators: OpenAI and hyperscalers creating AI models from scratch. These entities face unique regulatory challenges related to responsible and demonstrable data sourcing.
2. AI adapters: Fine-tuners of the creators’ models that embed them along with retrieval-augmented generation and similar technologies, adapting them for specific business application development. Enterprises must ensure they are sourcing models that can be attested to on intellectual property (IP) infringement or have some sort of protection against IP infringement.
3. AI consumers: Most businesses taking advantage of the adapters’ AI applications in their day-to-day operations. These organizations must ensure their data sets are cleansed and compliant with regulations.

The perils of AI missteps

We’ve seen a lot of enthusiasm from CIOs trying hundreds of different uses for generative AI. But two years in, they’re still largely lost on how to scale and monetize it.

Enterprises need to get back to the basics, treating AI initiatives like traditional IT application development rather than uncharted territory. Narrowing the scope to two or three projects can make innovation more manageable and bring demonstrable value. That means asking fundamental questions like: How do we vet the use case? How will we validate it? How will we support it once it’s built?

Start with a clear understanding of the business problem, create requirements, and ensure the solution will generate a measurable advantage, such as increased productivity or cost savings. It’s also crucial to follow a structured development lifecycle that includes cost controls, security measures, and governance.

Build from where you already have good utilization metrics. For example, you might run a call center and know that your customer service representatives handle 10 calls per hour. If you deploy a tool that allows them to handle 15 calls, that’s easily measurable. The key is to find opportunities within your organization that you can optimize and deliver a smarter process.

6 practical steps to AI

There are several ways in which your organization can establish a solid AI standard:

1. First look inward: Focus on internal applications to optimize workflows, automate processes, and reduce risks. They are easier to identify—and safer because you’re not exposing yourself to external vulnerabilities. Starting here also allows organizations the grace to learn and adapt before expanding to applications that impact more stakeholders, including customers. You can scale outward once you understand a solution’s full implications like security concerns, legal ramifications of how the models were trained, how they’re really licensed, and your operating costs.
2. Ensure data integrity and compliance: Data integrity and compliance are critical for all three AI use case categories. For creators, ensuring responsible sourcing of data is essential. Adapters need to cleanse and comply with data sets, while consumers must vet software-as-a-service providers and confirm proper data management.
3. Follow the lead: Learning from state-level regulations, such as California’s, can offer insights about future federal frameworks. Businesses should learn from how others adapt accordingly.
4. Adopt ethical AI: Implementing responsible practices is imperative to navigate the regulatory landscape. Business leaders and technologists should prioritize transparency, data privacy, compliance, and continuous learning in their AI programs, along with flexibility to adapt to new or changing regulations and technologies.
5. Surround yourself with knowledgeable teams: Leaders should surround themselves with knowledgeable teams to navigate AI’s complexities and understand their business’s true needs. AI projects’ success rely on a cooperative effort from cross-functional teams, including business functional areas addressing specific challenges, development, data science, IT, and FinOps. Establishing an AI center of excellence unites them.
6. Avoid past mistakes: The current rush to adopt AI mirrors past technology adoption cycles, such as the rush to adopt cloud services without proper planning. Avoid being swayed by the allure of new technology without assessing its implications. Instead, methodically approach AI as you would any other enterprise tool.

Our industry is at a leaping point from abstraction and conceptual thinking to tangible AI implementation. The goal is to find the real value in the challenges it can solve for your business.

For AI to generate new revenue streams and streamlined operations, prioritize practical solutions over grand innovations. Focus first on the unsexy work that frees your employees from the mundane tasks that no one loves.

Moving beyond merely trying AI to doing AI requires starting with sound processes and practical applications that not only will insulate your organization from future uncertainties, but drive it forward. Returning to IT fundamentals is the key to making AI a reality.

Juan Orlandini is CTO, North America of Insight Enterprises.

https://www.fastcompany.com/91280642/playing-by-the-rules-of-ai?partner=rss&utm_source=rss&utm_medium=feed&utm_campaign=rss+fastcompany&utm_content=rss