As told by Vishal Bhalla, SVP Enterprise Chief Experience Officer at Advocate Health and HR Exchange Network Advisory Board
Artificial intelligence (AI) and generative AI are forcing Human Resources leaders to pay attention and take charge. They have the opportunity to lead organizations as they adopt this advanced technology. However, not everyone in HR knows where to begin. HR Exchange Network recently sat down with Bhalla, who explains how to navigate adoption. He starts at the beginning:
Three Superpowers of AI in HR
First, AI can help humans become more efficient and productive. In this way, AI functions as a tool for the people. It enhances the skills and talents that employees already have.
It can replace humans, in part, to accomplish repetitive tasks. For example, chat bots can answer frequently asked questions using text or audio. Users can even create a friendly avatar to make it look more human and make the experience more inviting.
The third superpower is AI’s ability to do things that are not humanly possible. For example, HR might have much qualitative data from surveying 155,000 teammates in the last three years. AI can break down the results, share the demographic profile, determine what’s most important to each segment of the population, recognize trends, patterns, and more.
WATCH: 2024 AI in Hiring Trends
HR’s Role in the Revolution
HR can enable the adoption and adaptation to AI. To start, HR simply can help leaders rethink each job description. Parts of these functions could be understanding robotic process automation, machine learning, generic neural networks, or learned language models. HR can look at all the functions performed by a particular department to categorize the group’s tasks. Then, they can determine what tasks humans should do and what tasks AI should do. What will the new for a human look like? What new skills do people need? The next step is to upskill talent to meet the new criteria.
Leaders cannot stop at retraining humans. They also need to upskill the AI to represent the company to the customer, whether it is an internal or external one. For example, the voice responding to questions should not sound like a synthesizer. It should have the tonality, the language that reflects the culture of the organization.
HR should be able to enable preference by language and modality to reduce friction. HR has to then educate the customer, and explain the new ways of interacting with HR or the organization. Amazon and Netflix have done that well. It can’t be big brotherly. Right. That’s where HR can add value. HR can recognize the different perspectives and bridge that adoption.
The Challenges
The first hurdle is for HR to overcome its own preconceptions. They must work on their own hearts and minds and then those of leaders in the organization.
HR oversees training in many organizations. Now, it must consider training AI models, too. This will ensure that the model is not being trained on irrelevant sets of data in relevant rations, which would make the model bias. Human Resources must help with the training of the model. HR has to ensure ethics, compliance, and alignment with inclusion strategies. Users can actually leverage technology and neural networks to overcome the gaps that humans may have.
Maybe you’re serving a group that is 60%, Caucasian, 30%, African American, and 10% mix, which is made up demographics. HR can actually use the technology to fill in the data to normalize it to represent their population, which no human can do. As a result, HR can actually make the data unbiased, and then train and test the model against real samples.
The second huge hurdle that you have to overcome is adoption by leaders. There are two kinds of leaders – IT team leaders and operational leaders. How many humans are experts in AI, who HR can hire? They will never have enough. As a result, HR must have a two-way approach. First, take early adopters, and train them in no-code models. Set parameters within which they can pilot in low-risk scenarios. And a no-risk scenario is when people can open this lock to see what happens. If they don't open this lock, they see what happens. The human is always in the loop.
Finally, the point is that HR must get them comfortable enough to try it to address a pain point. For example, “I have to take calls for complaints.” Perfect. Let's take that away. How about you only do complimentary calls and not complaint calls? They must build trust for a period of time by demonstrating the value of taking away the non-value add and training them. There will be some of them who are early adopters. Train the early adopters on the technology. They can take on parts of the tasks that are low risk and train the model. Then, the peer-to-peer adoption rate will get generated and it has to be supported by the specialty SMEs.
Pondering the Future with AI
HR leaders must be at the strategic table when organizations are talking about AI. Foundationally HR must have data, which is clean. Also, the definitions have to be consistent, or HR will end up in trouble. Companies must have the ability to put everything in one data lake or data warehouse. So, data, cleanliness of the data, and the translational validity of the data or data governance is important. HR has to build a trust between the high skill SME and the operations SME who needs to adopt this wide skill. Most importantly, everyone must have acceptance of failure. Nothing will work in the first instance.
This technology improves at an exponential rate. Organizations will go from poor to suddenly starting to see the results. It's important to have strong fundamentals when it comes to the data and technology, then the skill, and then the culture in that sequence.
It’s all about humans leveraging it.