Enterprise AI Governance

What is Enterprise AI?

This technology combination enables enterprises to meet business goals by leveraging vast stores of enterprise data to drive operational improvements and to create net new revenue streams.

‍Enterprise AI is the combination of AI’s human-like learning and interaction abilities with custom-designed software catering to an enterprise's needs. When done effectively, the integration of AI into an enterprise’s technology stack improves processes, intelligence, and capabilities in a safe, ethical, and responsible way.

Corporations are betting big on the use of AI within the enterprise.  In 2022, global corporate AI spend reached $92 billion, a sixfold increase over the level of investment seen in 2016.  As Chat GPT wasn’t introduced until the end of 2022, the massive influx of investment into enterprise generative AI that followed isn’t reflected in these numbers.

Enterprise AI Adoption

Enterprises are leveraging AI algorithms, Machine Learning (ML) and deep learning techniques to address specific needs in industries like healthcare, investment management, banking, insurance, CPG, retail, and government.  Enterprises are addressing these use cases using both traditional and generative AI methods.

Up Leveling AI Model Development

As enterprises look to create an AI strategy, they will need to consider new requirements that may not have been as critical before.  Increasingly, AI models are becoming part of mission critical applications.  With this change, the teams that build AI models need to up level the way they build and deploy AI models.

For instance, teams building AI models are likely to have to take a close look at what is often called the “ilities”.  This refers to a whole series of non-functional requirements that end with “ity”

As the strategic importance of AI models increases, so will demands associated with ensuring that AI models conform with enterprise grade, non-functional requirements.

From AI Guidance to Regulation

The EU AI Act is the first set of regulations with the force of law specifically focused on AI use.  

The EU AI Act, seeks to codify into law the 2019 ethics guidelines for trustworthy AI that were developed by the AI High Level Expert Group (HLEG).

The EU AI Act takes a tiered approach to risk, recognizing three categories of AI uses:

1. those uses that present unacceptable levels of risk to society and should be banned outright;
2. those uses that present a high level of risk and should be subject to the highest levels of scrutiny and
3. those uses that present minimal or limited risk and thus require a lower level of oversight.

The Act applies to all companies, irrespective to where they are located, that offer AI products or services in the EU market.  So, like the EU’S General Data Protection Regulation, the impact of this new Act will be felt far beyond the borders of the European continent.

With so many AI use guidance documents being issued by so many governmental entities around the globe, it seems certain that more governments will follow the path taken in the EU - evolving guidance into AI specific regulations that will have the force of law.

‍

Stronger AI Governance - An Urgent Need

Another area where enterprises will have to uplevel capabilities is AI governance.  The increasingly pervasive use of AI models driving decisions that can impact the lives of customers and employees has governmental entities everywhere taking notice.  Businesses will need to ensure that they have the right governance processes in place to meet existing and future regulatory requirements.

The regulatory trend It is clear. The EU Artificial Intelligence Act, NIST AI-Risk Management Framework, US Executive Order 13960 and the California Attorney General Office’s AI/ML Governance Request to all California Healthcare Providers were all published within the last five years.  

The AI efforts of enterprises globally will increasingly be the focus of audits and potentially significant penalties for non-compliance.  

For instance, failure to comply with provisions of the EU AI Act as it relates what the act defines as high-risk AI uses, can result in fines up to 20 million euros or 4% of turnover.

The Consequences of Poor Governance

Though the fines associated with not meeting regulatory requirements can be substantial, the potential downsides of poor governance extend beyond regulatory compliance.  Good governance also helps to ensure that the AI models that underlie enterprise applications are valid and accurate.  

Validity means more than model accuracy.  Models must also prove that they achieve the business objectives set by enterprise.

The lack of an effective governance framework can create serious negative impacts for any enterprise. Consider the experience of Zillow, a real estate marketplace company.

In November 2021, Zillow announced that it would close “Zillow Offers” and cut 25% of the company’s workforce. The home-flipping unit’s challenges were the result of an unacceptably high error rate in its machine learning algorithm used to predict home prices.

Many factors contributed to the poor performance of the algorithm Zillow used to predict home prices but if AI is used to drive significant revenue enterprises absolutely need effective governance in place.  Effective governance helps to ensure that the right people and teams sign off on models before they go into production.  

Enterprise Grade Governance

If enterprises are to succeed at ensuring AI use accelerates innovation while at the same time safeguarding the organization against increased operational and regulatory risk, they will need to modernize their approach to AI governance so that it represents a true enterprise grade capability.

Key attributes that need to be part of a modernized enterprise AI governance framework include:

• Comprehensive Inventory - Visibility into all your AI initiatives with a dynamic inventory that integrates with your priority AI systems
• Light Controls - Implement a risk-based compliance approach and enforce the requisite controls for all AI systems
• Robust Reporting - On demand reporting into AI usage, risks, and adherence across internally developed, proprietary, vendor, and embedded AI

‍