Williams Lea CEO Clare Hart discusses how HR can embrace AI and thrive
Will AI deskill HR? A study released by Harvard Business School found that rather than replacing human employees, AI will create a “jagged technological frontier” due to technology such as large language models (LLMs) performing some tasks perfectly while others need human intervention.
The study monitored how 758 Boston Consulting Group consultants integrated LLM tools into their regular work processes, and found that “some tasks are easily done by AI, while others, though seemingly similar in difficulty level, are outside the current capability of AI,” researchers wrote.
Clare Hart, CEO of Williams Lea, has seen firsthand how AI integration into a vast workforce can be used to enhance skills rather than replace them. With the implementation of AI, she explained, it will become more important for there to be a “human in the loop” to conduct quality assurance at a level only capable by humans.
Entirely new roles will be created to fill in gaps where AI can’t reach – especially in the case of generative AI such as ChatGPT, Hart explained, and human intervention will be necessary.
“The technology could misinterpret things because it doesn't have the EQ that we do. It doesn't have the human experience to know how to filter, and it just brings things in that don't make sense,” she said. “The human QA is going to be critical to the successful rollout of artificial intelligence in in the business community.”
Prompt engineers among newly created AI roles
Because of inherent errors or inconsistencies in AI-generated content, there will need to be “prompt engineers” to supplement AI and ensure the content is clean, Hart said.
“The prompt engineer is the one that helps build the guard rails,” she said.
“They have technical skills, and they're going to have to understand the framework of the large language model … Is it about litigation around clean energy, or is it about litigation more broadly? They have to build the framework so that you're not looking at the whole web, or you're not looking at everything in the company's database, but you're looking at the information around this industry or this topic.”
Prompt engineers can also assist with issues of bias that are emerging as main challenges for AI in HR, said Hart.
“A lot of testing has to be done on these new generative AI models, because you have to test it to make sure that those things don't happen. And then if they do, refine the coding and the technology to eliminate the risk of that unconscious bias.”
AI is not a threat to HR jobs, but more skilled people will be
The Harvard study found that individual employees adopted AI tools to various degrees, but all of the consultants who used AI showed improved quality of work (40% improvement over control group). Significantly, employees who performed below the average threshold improved their work quality by 40%.
Hart stressed that the perception that AI will deskill and replace human employees is incorrect. In fact, it is other employees who will be doing the displacing, if AI isn’t fully embraced.
“AI is not going to take your job,” she said, “but somebody who understands AI is going to take your job. Somebody that understands how to use AI, may take your job.”
“Everybody has to stay in that continuous learning mode, and continue to figure out how to use AI, on your enterprise apps or even in the consumer apps like ChatGPT and Bard, and get out there and use it.”