Skip to content
Fast Company Executive Board

The Fast Company Executive Board is a private, fee-based network of influential leaders, experts, executives, and entrepreneurs who share their insights with our audience.

AI guardrails: Will they protect or paralyze?

AI is here to raise us up, not replace us altogether. But guardrails are still needed.

Profile photo of Amit Walia

CEO · Informatica

The concept of artificial intelligence (AI) has been a prominent part of our culture for decades, with films like 2001: A Space Odyssey and The Terminator,igniting our imaginations to the possibilities of this not-so-distant, yet still unobtainable, technological revelation. As entertaining as these works are, there’s often an underlying theme that permeates the relationship between humans and AI—inevitably, something goes wrong. 

As AI/Generative AI (GenAI) has exploded in the past 12-plus months, perhaps that’s why we have witnessed just as much extreme pessimism and fear as we have optimism and excitement. For the first time, all of us, from technophiles to technophobes and everyone in between, can experience this innovation at our fingertips. 

While it may not be something as dystopian as rogue AI attempting to harvest humankind, the trepidation to these AI advancements has largely been around our collective run-before-we-walk mentality, and the possibility of unforeseen consequences to job markets, personal safety and privacy, intellectual property, and corporate security and governance. AI was prominent on the picket lines last year as actors and writers sought creative protection, and it’s already become a central theme in global elections this year as many parties look to wield disinformation to influence ballots.

During a conversation I had with James Maguire, editor-in-chief at eWeek, he described an almost quasi-hysteria since ChatGPT launched at the end of 2022 with two entrenched camps: one calling for AI/GenAI’s unbridled liberation, and the other, its immediate curtailment. The truth is, as we witnessed with the emergence of cloud and distributed computing, with something this transformative, both reactions are expected and appropriate.  

Here is the paradox: In the short term, I believe there’s more hype than reality around AI/GenAI, but in the long term, we haven’t even begun to dream of its true potential. Its promise of profound value creation has every business leader feverishly exploring its undeniable benefits to their operations. At the same time, many continue to fumble in the dark to fully understand how they scale, operationalize, and manage this technology appropriately, including the moral, ethical, and soon-to-be complex legal responsibilities that come with it. 

In a study we commissioned with 600 global data leaders, respondents said data privacy and governance (40%) and AI ethics (38%) were key hurdles to GenAI adoption. Additionally, 39% of respondents listed the ability to deliver reliable and consistent data fit for GenAI and improving data-driven culture and data literacy as top data strategy priorities this year.

AI GUARDRAILS ARE INEVITABLE: GOOD FOR BUSINESS OR BAD FOR INNOVATION? 

Make no mistake, AI/GenAI regulatory compliance is only a matter of time. It may be a while before there’s consensus on legislation, but the days of Wild West-like freedom are numbered.  

As AI/GenAI continue to scale intelligence, unleash productivity, and democratize complex tasks previously impossible for average users, guardrails will be established to ensure the technology not only functions safely and properly, but that users don’t intentionally or unintentionally abuse its power. 

We’ve seen a growing appetite from policymakers around the world to rein in this innovation and codify formal AI frameworks, including here in the U.S. with President Biden’s Executive Order and various state legislations, as well as the European Parliament’s Artificial Intelligence Act. As a technologist and early adopter of AI into our own business and product solutions, I embrace the philosophy of “letting a thousand flowers bloom.” We must be thoughtful and collaborative on what needs to be regulated and how this regulation is enforced; if the leash is pulled too quickly or forcefully, it can kill the momentum that’s driving much of the progress we’ve recently witnessed. It’s imperative lawmakers not unilaterally erect these guardrails without meaningful consultation with software/AI companies, business leaders, advocacy groups, and other key stakeholders. 

But what exactly will these guardrails look like? Expect this to be twofold: The first will be regulatory compliance, established by international coalitions, nation-states, and local state/provincial jurisdictions. These comprehensive laws, such as the EU’s Artificial Intelligence Act, will look to close loopholes, establish regional (or global) standards, and protect the public from broader misuse, such as the loss or mishandling of data or bad actors who leverage AI/GenAI to steal information and compromise security.  

The second will be internal policies unique to each organization and crafted by cross-departmental teams or working groups. These guides, which can and should be developed and implemented now, will protect the reputation of an organization and the hard-earned trust established with its customers. Businesses should provide clear expectations for employees, partners, and contractors on everything from AI/GenAI creation, experimentation, and deployment and bias awareness to who can access and manage certain data sets or AI/GenAI applications.  

DON’T FORGET THE DATA 

With all the hype around AI/GenAI, what often gets lost is what’s under the hood: data. Trusted data is the crux of any successful AI/GenAI project. Unfortunately, a persistent problem among organizations is a lack of understanding or confidence in the quality of their proprietary data. 

This is why proper controls throughout fragmented and dispersed data ecosystems are paramount; without them, organizations may amplify incorrect insights, or worse, base key business decisions on outputs generated from poor underlying information. Currently, and in the face of impending legislation, investments in data quality, governance, and security are imperative to successful AI/GenAI workloads, as well as broader organizational health, growth, and stability. 

Plainly, with AI/GenAI, we don’t know what we don’t know. Every other day, there’s a new development or breakthrough. We’re on this rollercoaster together. Despite understandable uncertainty over the future of AI/GenAI, I’m incredibly bullish on its reward over risk, its potential over pitfall, and the myriad of use cases we’ve yet to comprehend that will elevate civilization.  

AI/GenAI is here to raise us up, not replace us altogether. Regulations and guardrails will be a foundational part of that conversation, and as business leaders, we shouldn’t shy away from them. Instead, acknowledge their approach, participate in the dialogue, educate our teams, and embrace the revolution. 


Amit Walia is the Chief Executive Officer of Informatica, an Enterprise Cloud Data Management leader.