AI in the Factory: What Manufacturers Must Consider

As generative artificial intelligence and machine learning continue their rapid rise in popularity and uses in business operations, manufacturers are faced with how to implement AI in a manner that allows them to take advantage of the technology while remaining mindful of the associated risks.  

To responsibly implement AI into business operations, manufacturers should consider the following action steps.  

Carefully Select AI Programs for Use 

When selecting which AI programs to implement, manufacturers should carefully vet the program before purchasing and/or contracting for those programs to ensure, to the extent possible, that the programs are well-developed, routinely updated, contain data privacy protections, and have built in safeguards for other inherent risks presented by the use of AI. 

Implement Data Governance and Access Controls 

AI is data driven and relies on the input of data—some of which may be sensitive—to function.  

Because the use of AI will require the input of company data, manufacturers should take steps to protect the data that is input into AI programs by: 

• Classifying data to ensure a firm understanding of what data is being used and how 

• Controlling access to data using role-based permissions or other access control procedures 

• Implement logging and auditing for AI program accessing or data ingestion to ensure the ability to identify and rectify any potential breaches 

Take Intellectual Property into Consideration 

Intellectual property must be considered at the ingestion and output stages of using AI models.  

On ingest, manufacturers should ensure that sensitive data, such as trade secrets, are either not ingested into the AI program, or that the agreements with the AI suppliers are crafted to ensure the protection of trade secrets.  

At the output stage, manufacturers should carefully review and/or negotiate contracts with AI suppliers to develop a firm understanding of who owns AI outputs. 

Consider Vendor, Supply Chain Risks 

Manufacturers should consider how vendors and suppliers are using AI and whether their use can have up or downstream impacts.  

Manufacturers should review software licenses and SaaS terms for IP clauses and indemnities, conduct vendor security audits, and ensure data localization or export restrictions are met to mitigate risk. 

Review Insurance Coverage 

Review existing insurance policies for potential exclusions for losses resulting from the use of AI.  

Consider where gaps may exist between policies: for example, professional liability policies may exclude AI-related injuries, but cyber policies may not cover other losses typically covered by those policies.

Implement Robust AI Governance Policies, Educate Employees 

Any use of AI in the factory may initially be run by employees on a day-to-day basis.  

As manufacturers put care into selecting and implementing AI programs, they must also ensure that the policies governing the use of those AI programs are robust enough to address the steps manufacturers are taking to mitigate risk, and that employees are provided with the education and training necessary to use the tool effectively and ensure adherence to the adopted AI governance policies.  

Takeaway 

The rapid rise of AI has the potential to revolutionize manufacturing, but manufacturers must take the appropriate steps to address the risks presented by AI and machine learning to ensure that the benefits are not outweighed by the risks.  

This article first appeared on CBIA's website and is published here with permission.