A recent study by the Boston Consulting Group (BCG) found that despite headlines warning that artificial intelligence poses serious risks to society, employees are showing curiosity, optimism and confidence in the arrival of AI in the enterprise. has, and has become even more pronounced over time.
For many people, that feeling is based on their experience of sound. ChatGPT, a representative of generative AI applications, was just released in November 2022, but already 26% of workers use generative AI several times a week, and 46% generate it at least once. It was discovered by BCG that it has experimented with AI.
BCG asked 12,898 frontline employees, managers, and leaders from large organizations around the world how they feel about AI. 61% said curiosity was the two strongest emotions, 52% said optimism, 30% worried and 26% said confidence. In his small BCG survey five years ago, 60% were curious, 35% were optimistic, 40% were concerned, and 17% were confident. A lot has happened since the last survey of AI attitudes in 2018. At the time, “AI was still emerging and not something that many people were using or seeing,” said Vinciane Beauchène, BCG’s Global Talent and Skills Leader. said at the conference. Make a phone call to discuss the findings. But after the pandemic rocked the world of work, generative AI has flourished. “This is the new normal. Her 80% of leaders claim to take advantage of it every week,” she said. “What strikes me is that companies are still not ready to address this issue.”
Familiarity creates content
The study also found that more workers using AI tools became less concerned and more optimistic about their impact. He was concerned by 22% of normal AI users and 27% of rare users, compared to 42% of non-users. Meanwhile, 36% of non-users said they were optimistic about AI, compared to 55% of rare users and 62% of regular users.
Companies can help employees move from fear to optimism, said Nicola de Belfonds, global AI leader at BCG X, the company’s technology building and design arm.
“What we have seen, and over the past eight years, helping companies make the most of AI, is how employee adoption of AI can improve and improve their jobs. It’s directly tied to employees’ understanding of what it means to be empowered,” he said. .
We still have a lot to do with this score. So far, only 14% of frontline workers and 44% of leaders have received AI training, yet 86% of survey respondents believe AI will be needed in the future.
“That’s a huge number,” says Belfonds. “We really have to deal with this upskilling problem.”
Change management
However, it is difficult to get employees to understand what AI can do for them, as they want to feel that AI is augmenting their work rather than just replacing the fun part. BCG found.
“The hardest part of embracing AI is creating a space where employees can add value without feeling like they are competing with AI to create value,” Belfonds added. “A lot of the work we do around change management and coaching is about helping people take advantage of AI while also changing the way they add value. Some will be taken over by AI, but some of our employees’ jobs will be taken over by AI.” Refocus on higher value-added tasks. ”
How these processes are rewired and how they work will be different for each company, he said.
There are other ways that employee concerns about AI are unevenly distributed. BCG found that leaders tend to be optimistic and frontline employees are concerned. And while 68% of his leaders believe his company is taking appropriate steps to ensure the responsible use of AI, only his Only 29%.
Despite the optimism in the workforce, there is a dark side, according to BCG’s findings. More than a third of his respondents believe their jobs are likely to be displaced by AI, and almost four-fifths believe that governments should ensure that AI is used responsibly. Hoping to intervene and introduce his AI-specific regulations. The percentages were highest in India (89%), Spain (88%), Italy (84%), Brazil and France (both 83%), and lowest in Japan (64%) and Germany ( 73%), USA (74%), Middle East and Netherlands (both 76%).
5 step program
While waiting for regulators to decide what can and can’t be done, Stephen Mills, BCG’s chief AI ethics officer, asked CIOs a few questions about how to safely introduce generative AI into the workplace. I am proposing.
Familiarity with generative AI is a key factor in a successful implementation, so employees should get the chance to test it themselves.
“It’s important that people have the opportunity to interact with and use these technologies. Stopping experimentation is not the answer,” Mills said, noting that it’s also not realistic. “AI will be developed across organizations whether employees know it or not,” he added.
Instead, he suggested: “Instead of trying to pretend it doesn’t happen, let employees know where the guardrails are, what they can and can’t do, and put in place a set of simple guidelines that actually encourage responsible innovation. and experiment responsibly.”
Investing in upskilling (ongoing rather than one-time) can also help, especially among frontline workers with less tech savvy. “This disparity between levels will be very important for businesses to understand as they consider this ongoing journey of change that is underway … equipping businesses with the right skills to succeed as an organization. for,” he said.
Finally, he advised companies to build responsible AI programs to reassure employees that they are using generative AI ethically.
In summary, Mills said there are five key elements to such a program. An overarching principle that defines a company’s strategy and risk tolerance. A governance structure that defines the organization and escalation paths. The process of integrating AI into product development. The tools you need to do all this. and a driving force for cultural change.
Specifically, there are some immediate steps companies can take, including delegating responsibility and accountability for the program to senior executives, he said. Ensure that the person and their team have the funds and resources to build the program. Then immediately install the initial guardrails and ensure they are adhered to.
That means “having an agile review process where teams experimenting with AI can work together and reach out to answer questions,” he said.
