8 challenges of implementing AI in L&D

1.     Data integrity

AI needs clean data to be efficient and effective. For example, predictive analytics requires a lot of good quality data to ensure the predictions are accurate. Low-quality data could cause unreliable forecasts, and analytics are useless if not enough data is present.

Differentiating between historical data and new data is important in case data integrity issues occur. Issues with historical data can be fixed with a one-time cleanup. However, if new data is problematic, fixing broken processes that are causing data issues will be necessary before data cleaning occurs.

2.     System integrations

AI might need to draw on data from multiple systems to carry out tasks such as personalized training recommendations or predictive analytics. The data from all HR systems should be integrated into one database if the HR team uses multiple vendors for systems such as the HRIS, talent acquisition software, learning management software and performance management software.

However, the HR team might not be able to easily merge the data without investing in a data management tool, which could result in additional expenses and increased HR system landscape complexity.

3.     User knowledge

HR employees might struggle with deciding how best to use AI for learning and development, so CHROs might want to consider bringing in an expert consultant. They could create a new role tasked with integrating AI into HR processes and training the HR team or expand on an existing one.

4.     Vendor features

In the short term, HR teams might be limited in what they can add for AI L&D capabilities because of the company's current vendors.

Those vendors might already offer their clients AI functionality and share a vision for how they will integrate AI features in the coming years. However, if they aren’t providing enough AI functionality, CHROs might want to switch to another vendor.

Moving from one vendor to another could result in long-term disruption. Also, the company might discover later that the promised AI features are not fully developed or do not work as well as they appeared to during the demo.

5.     Cost-benefit analysis

Some AI functionality is low-cost, including online applications that offer writing suggestions. However, HR systems with AI functionality could require a large investment.

CHROs must understand the expenses that a new system will require compared with potential efficiency-related savings before they commit to a new application. Existing processes might be cumbersome, but spending money to streamline and accelerate them with a new AI tool might not be worth it.

6.     Capacity shortfall

Some HR teams have small L&D teams, so while L&D employees might be interested in improving processes with AI, they might also feel that they lack sufficient time for AI experimentation.

Therefore, the CHRO might need to increase the team size or use IT resources for AI experimentation for a brief period. They might also need to cut back on other projects so employees can focus on AI.

7.     Data privacy

As AI gains more access to employee data, clear boundaries must be set and continually tested.

For example, an employee using AI to summarize their own employment history is fine, but AI should not be permitted to do so. Also, AI should not be permitted to access confidential employee data, such as an employee's Social Security number, race or marital status, when making recommendations about workers.

8.     AI errors and embellishments

L&D employees will need to be on guard for AI hallucinations referencing made-up facts. For example, AI could cite a legal case that doesn't exist.

HR team members will need to review any AI recommendations or other work, so CHROs should consider whether adding AI capabilities will result in time savings.

Adding AI capabilities to L&D can help companies in some ways, but CHROs need to confirm that the benefits of the technology outweigh challenges related to cost, integration, complexity and security. Keeping the most common AI L&D tradeoffs on their radar can help organizations avoid bad investments.

Eric St-Jean is an independent consultant with a particular focus on HR technology, project management and Microsoft Excel training and automation. He writes about numerous business and technology areas.