What US Manufacturers Need to Know about AI Regulatory Compliance

With the growth and widespread application of artificial intelligence (AI), there is a distinct need for regulatory compliance to ensure efficient functioning and protect all stakeholders. Lori Witzel, director of research for analytics and data management at TIBCO, discusses what all US manufacturers must stay updated about in AI regulatory compliance to gain and maintain a competitive advantage.

There are high-value artificial intelligence (AI) use cases that can provide manufacturers with significant competitive advantages, including advanced asset management, predictive maintenance, and anomaly detection. However, US manufacturers need to be aware of the potential impact of increasing AI regulation. 

The Business Value of AI for Manufacturers

For manufacturers, AI and machine learning (ML) can provide predictive, prescriptive, automated insights and actions across various use cases with various business benefits. AI can improve outcomes when operational decision-making requires more agility through insight, scale, and speed in trend investigation, anomaly detection, root cause analysis, or key factor identification. 

Here are some specific applications of AI in manufacturing; every manufacturer should focus on which data-driven innovations and efficiencies would be most impactful for their business.

• Yield optimization via pattern recognition at scale: Visual data can provide insights into manufacturing defects that affect yield, but turning the enormous volumes of data from visual streams and sensors into actionable insights requires AI / ML applied to pattern recognition and anomaly detection. Hemlock Semiconductor is an example of a manufacturer using this AI approach to transform its business and open new markets. The agility provided also helped Hemlock maximize its resource efficiency while speeding mean-time-to-resolution on potential quality issues. 
• Predictive and prescriptive maintenance via advanced equipment monitoring: Equipment condition monitoring using AI / ML on sensor or telemetry data automates the analysis of that data—which may be variations in appearance, vibration, temperature, or noise—to predict failure and prescribe remedial actions, such as ordering replacement parts or scheduling maintenance. One real-world case is how Brembo’s manufacturing department uses AI-infused analytics for predictive maintenance and lifespan prediction of machine tools.
• Asset management optimization via digital twins: Digital twins are an AI-infused virtual representation of an as-is physical product, with its systems and related processes, using real-time process data and analytics derived from precise configurations of the subject of the twinning. For manufacturers, digital twins enable more efficient equipment maintenance and optimize designing, assembling, deploying, and testing production systems. IDC predicts Opens a new window that by 2024, fully half of G2000 manufacturers will develop ecosystem digital ops centers – including digital twinning – to build capacity (both human and machine) that will yield 50% faster time to market.

Those manufacturers that lag their peers in adopting AI-infused technologies are now at a competitive disadvantage. Research from McKinseyOpens a new window shows leaders improved performance across a number of KPIs by 10% or more compared to lagging peers. With an overall average performance improvement of 9.5%, leaders had nearly three times the improvements gained by laggards.

See More: A Quick Guide to Smart Manufacturing

Current Situation: AI Adoption and Growth are Driving Regulation

According to recent researchOpens a new window by the Harris Poll and Google Cloud, 64% of manufacturers use AI in daily operations, with roughly 25% using 50%+ of their overall IT spend towards AI. While it’s clear the time is now for manufacturers to take advantage of AI / ML, there are legal and ESG dimensions to consider. 

AI is a hot topic among the public, often associated with threats to job security and social displacement. Due to the potential for negative as well as positive impacts from AI – the same technologies that provide manufacturers value can also reinforce inequity – legislators, policymakers, and government agencies are considering AI regulations to reduce the risk of harm.

As manufacturers seek value from AI, building trust and transparency into AI is a core best practice. It’s also imperative to ensure compliance with current and future regulations.

Trustworthy, transparent AI will reduce the risk of serious harm resulting from flawed AI. The potential for serious harm and public outcry over those harms leads governments and policymakers to develop AI regulation. There are additional benefits beyond risk mitigation from ensuring AI is transparent and auditable. Auditability enables processes to be more easily reused/cloned once AI has been shown to add value, helping manufacturers accelerate their digital transformation.

Details on US AI Regulation and Key Takeaways 

Below are recent developments related to AI regulation in the US. Note that the EU is also developing AI regulation. If you have customers, suppliers, or partners around the globe, you’ll need to follow those developments. In part two of this series, we’ll discuss the steps manufacturers can take to prepare for new and current regulations. 

In the United States, specific AI regulatory guidelines have been proposed on an agency-by-agency basis at both Federal and state levels. The Federal Trade Commission’s April 2021 blogOpens a new window , “Aiming for truth, fairness, and equity in your company’s use of AI,” indicates the FTC will use its authority to pursue the use of biased algorithms. The FTC’s ongoing enforcement actions show they’re serious about pursuing compliance for AI fairness and transparency. As they wrote, it’s time to “hold yourself accountable – or be ready for the FTC to do it for you.” 

While it’s still in development, the National Institute of Standards and 3 Technology (NIST) Artificial Intelligence Risk Management Framework (AI RMF or framework) will address risks in the design, development, use, and evaluation of AI products, services, and systems.  

Although AI regulation is not yet law for a majority of all states, momentum toward regulation continues. As of this date, states with AI regulations enacted include Alabama, Colorado, Illinois and Mississippi. States with AI regulation pending include California, Hawaii, Massachusetts, Michigan, New Jersey, New York, North Carolina, Vermont and Washington.

Although AI regulation Opens a new window is not yet law for a majority of US states, momentum towards regulation continues.

See More: 3 Big Data Challenges for Manufacturers and How to Solve Them

Let’s take a look at key takeaways for the years to come:

• Don’t focus on AI regulation in a single state or US region; seek the broadest possible approach to AI regulatory preparation: It’s unlikely a manufacturer’s legal responsibility will be limited to a physical headquarters or plant site. Now is the time to think globally in your planning. For example, an EU citizen residing in the US who is a partner, customer, or supplier of a US-based manufacturer means that EU regulations will apply.
• The US regulatory environment for AI is evolving rapidly, US manufacturers must keep pace: In many cases, although AI is a technological approach to optimizing manufacturing processes, yields, and outcomes, you’ll need guidance from legal consultants with expertise in these issues. Addressing the need for transparency and auditability is not just for your IT and analytics teams.
• Don’t assume that your technology stack is free from non-compliant AI; your vendors and suppliers may have AI in their own solutions: Even if your technology teams ensure your own company’s use of AI is transparent and auditable, the use of AI by your vendors and suppliers may pose a risk. Include their solutions in your review processes.

What’s next? Since artificial intelligence can provide breakthrough advantages for manufacturers – nearly a 10% performance improvement, according to McKinsey – its adoption will continue, as likely will accompanying regulation. In the second part of this series, you’ll learn five steps to take now to reduce regulatory risk as you pursue your newfound AI advantage.

Are you paying attention to AI regulation? How are you ensuring you’re compliant? Tell us on FacebookOpens a new window , TwitterOpens a new window , and LinkedInOpens a new window .

MORE ON MANUFACTURING

• How AR/VR Is Driving Automation and Robotics for Manufacturers
• Impact of Digital Transformation and Evolving Tech on Manufacturing
• How to Increase Productivity in Manufacturing Operations with Digital Technology
• 3 Ways Manufacturers Can Manage the Data Tsunami and Amplify Their Investments in IoT
• How to Build Smart Businesses (and Not Just Smart Factories)