Why governance should be a crucial component of your 2021 ML strategy
One of the key trends we discovered is that ML governance is by far the top challenge that organizations are facing with model deployment as they head into 2021. This post explores what we discovered in the data—and what it means for your organization as you head into 2021.
56% of organizations struggle with governance, and 67% must comply with multiple regulations
In the survey, we asked respondents to indicate the challenges they face when deploying ML models. While we saw a range of issues reported, the #1 challenge selected by respondents was IT governance, security, and auditability requirements. In fact, 56% of all respondents indicated that this was a challenge at their organization—more than any other challenge in the survey.
We also asked respondents about the regulations they need to comply with for their AI/ML efforts. The vast majority of respondents—67%—selected multiple regulations, while 26% reported needing to comply with one regulation, and only 8% selected no regulations at all.
What does it all mean? Though more and more organizations are investing in AI/ML, we’re still in the very early days of governance. Many organizations are (understandably) struggling with this issue in a complex and ever-changing regulatory landscape.
Why is governance important?
AI/ML model governance is the overall process for how an organization controls access, implements policy, and tracks activity for models. It’s a must-have to minimize organizational risk in the event of an audit—but that’s not all.
Governance is the bedrock for minimizing risk to both an organization’s bottom line and to its brand. Organizations with effective AI/ML governance not only have a fine-grained level of control and visibility into how models operate in production, but they unlock operational efficiencies by integrating AI/ML governance policies with the rest of their IT policies.
With governance, organizations can understand all the variables that might affect model results, which helps them quickly identify and mitigate issues (such as model drift) that can degrade the accuracy of results and the performance of applications. These issues can directly impact the business’ bottom line and erode customer trust in the brand over time.
What this means for your 2021 ML strategy
While 56% of respondents indicated that governance was a challenge for them, we suspect the actual number could be even higher. Governance issues tend to arise late in the ML lifecycle, after models have been developed and organizations are beginning to look for ways to minimize their risk. Many organizations could have governance challenges before they realize it.
Even if you’re not actively struggling with governance right now, it’s possible that it could crop up as a concern in the future. As organizations head into 2021, they should consider focusing more attention on governance even if it’s not currently a top concern.
Discover all 10 enterprise ML trends
This is only one of the trends that we discovered in our survey of 400+ business leaders. Check out the full report today to explore all 10 trends we uncovered in our research. 2021 will be a crucial year for ML initiatives—set yourself up for success by understanding where the industry is headed and how you can make the most of it.
More from the ML trends blog series
- ML trend: Enterprises are dialing up their machine learning investments for 2021
- Why an MLOps solution can accelerate your business in 2021
- Why you should pay off your technical debt for machine learning in 2021
- ML trend: I&O leaders are the most common decision-makers in cross-functional ML initiatives
- New report: Discover the top 10 trends in enterprise machine learning for 2021
More from the AI/ML governance blog series
- What is model governance?
- Why you need machine learning governance: Governor Brainard’s recent speech about responsible AI
- What you need to know about model risk management
- Why risk managers need to improve governance of AI in 2021
- Model drift and ensuring a healthy machine learning lifecycle
- The value of model accuracy
- Introduction to optimizers
- Introduction to loss functions
- How to version control your production machine learning models