Opinion: AI needs rules and standards. Otherwise, the risk may be greater than the profitability.

The promise of profound new types of productivity and unimaginable intelligence is significant, but with that potential comes unprecedented fear.

Leaders face uncertainty about how to adapt, optimize, and ultimately harness the power of technology. I believe we need to make AI-driven innovations happen and get them into the hands of as many people as possible, unlocking capabilities we’ve never seen or imagined before. But equally, businesses need guidance on how to experiment, test, and adopt AI ethically and safely.

Below is an AI governance roadmap to help leaders guide their efforts through 2024 and beyond.

1. Trustworthy data: In my world, so-called unstructured data needs to be cleaned, cleaned and sorted to capture it into chunks of business intelligence. But AI is only as good as the data that powers it, so the first step to strong AI governance is ensuring that the data itself is trustworthy.

According to a recent survey conducted by Eckerson Group, 46% of data leaders say their organization lacks appropriate data quality and data governance controls. Despite this, most business leaders don’t strive for relevance. They strive for excellence.

What does it look like? High-quality data is accurate, complete, consistent, timely, valid, and unique. For example, in the healthcare industry, it is essential to have complete, accurate, and unique patient records without duplication to ensure appropriate treatment, monitoring, and billing. A single mistake in managing that data can have costly and even destructive consequences, and this is before large-scale language models (LLMs) for AI or generative AI are built on top of it.

The problem, of course, is that most organizations still have highly fragmented data and lack a comprehensive governance framework. A natural starting point is assessing the current state of your data (what you have, where it is, how it moves, and how it is protected) to diagnose quality issues. Then apply rules to manage and monitor them to ensure quality is maintained over time.

2. Modern governance: Data governance is not a new concept, but it is becoming increasingly important due to the proliferation of AI and generative AI applications, stakeholder and shareholder demands, and global data regulations.

Many leaders have now outlined data governance practices that encompass people, processes, and technology across their organizations to create common rules and guidelines for collecting, storing, and using data. Many of these frameworks are based on risk and compliance, and for good reason. Protecting and using data ethically is an essential part of a successful data strategy that meets current and future data and AI regulations. For example, U.S. President Joe Biden’s recent executive order on AI will strengthen reporting standards for companies and government oversight of AI, and the EU is poised to announce AI regulatory and enforcement practices soon. there is.

Nonetheless, enterprises’ traditional governance approaches will prove inadequate when it comes to ensuring secure and effective data management for AI and generative AI. why? In most cases, they are not designed for the scale or democratization of data use that enterprises require today. The volume, variety, and velocity of data are too great for humans to manage.

Modern data governance is built on the table stakes capabilities of security and compliance to incorporate automation, adaptability, and agility. It increases productivity and results by considering and automating data integration, cataloging, management, and observability. In effect, this is not only about compliance, but also about exercising a competitive advantage.

3. Getting started: As leaders finalize their organizations’ 2024 strategies, AI will be a top priority. This is driven in part by democratization and increased access to generative AI. In fact, McKinsey reported that advances in generative AI will lead to 40% of organizations investing more in AI overall. But the smartest companies will prioritize AI governance to maximize value and minimize risk.

To get started, assess the current quality of your data and any underlying issues or disconnects in your data management processes. Build a collaborative team to develop and oversee governance practices to ensure the ongoing accuracy and reliability of data and AI-based insights.

Finally, keep repeating. Just as feedback is essential to product development and improvement, it is equally important to the performance of AI models. Your approach must be flexible to take into account new innovations and ways to use technology, and more importantly, new security threats that emerge every day.

Although we have never seen anything like AI, history is filled with countless innovations and opportunities that have inspired fear and forced us to build guardrails, from the gold rush and transportation boom of the mid-1800s to the communications and pharmaceutical developments of the 2020s. .^Day century.

History has shown time and time again that we can and will rise to challenges through the public and private sectors. The same will be true for AI if we work together to provide technology to as many people as possible and thereby secure their future.

Amit Walia is the CEO of Informatica, an enterprise cloud data management platform company.

more: We need a strong code of ethics to prevent the dark side of AI from overtaking us.

Also read: AI may not replace your job, but it will re-create your job. One thing should not change.