What HR pros need to know about AI in the workplace

Employees have enthusiastically embraced generative AI, with 91% of survey respondents saying they use it for work through internal or publicly available tools, according to research from management consulting firm McKinsey & Co.

Organizations, meanwhile, lag behind in their use of the technology, McKinsey found. To capitalize on employee momentum, companies must take a holistic approach to transforming how they work with generative AI, researchers Charlotte Relyea, Dana Maor, Sandra Durth and Jan Bouly suggested in their analysis of the findings.

Employee mindsets and behaviors are key, the researchers said. To win them over and drive successful change, the researchers recommended that companies follow four tenets:

• Role modeling, where leaders visibly adopt generative AI in their work;
• Fostering understanding and conviction through internal communications;
• Building capabilities, which involves comprehensive training on skills such as data analysis, machine learning algorithms and understanding AI-generated outcomes; and
• Reinforcing new ways of working by integrating AI goals into performance metrics and evaluation processes.

Right now, employee use of generative AI is at an inflection point, the researchers stressed. Organizations must be ready for the next inflection point — moving from individual experimentation to strategically capturing the technology’s value, they said. Otherwise, employers risk missing out on generative AI’s potential benefits and will fall further behind, the researchers emphasized.

McKinsey acknowledged that due to the rapid automation of many work activities, organizations are under pressure “to understand their talent and skill needs quickly, adopt various strategies to close skills gaps and invest in upskilling and reskilling.”

However, to create value — which technology alone can’t do, the researchers told business leaders — companies must reimagine talent and skilling.

Many are “buying” talent externally rather than invest in existing employees to build from within, according to an April report from The Adecco Group.

The firm’s CEO doesn’t recommend this as a primary strategy. While that’s one approach, it shouldn’t be the only one, he said in statement.

McKinsey researchers are more pointed: “A gen AI-based talent transformation isn’t something companies can simply hire their way out of, as it affects the entire organization and its ways of working,” they stressed.

Also, in identifying who’s right for learning new skills to work with AI, companies shouldn’t focus on job candidates or employees solely based on technical know-how, a Verizon talent acquisition executive recently told HR Dive.

Rather, HR managers should look for behavioral elements, such as having the ability to learn and coachability, which companies can “adapt to using these new ways of working,” the exec advised.

U.S. employers in an annual Littler Mendelson survey exhibited a nearly 50-50 split on the use of generative artificial intelligence for HR tasks, as 51% said their organizations did not use generative AI for HR while 49% did, according to results.

With respect to using AI for HR, most of the more than 400 in-house lawyers, executives and HR professionals surveyed were at least moderately concerned about complying with state, federal and international data protection and information security laws. That contingent included 81% of respondents representing large organizations, Littler said.

Employers were also concerned about employment law enforcement generally, with 8 in 10 stating they expected at least a moderate impact from the U.S. Department of Labor in the next year. Seventy-three percent said the same of the National Labor Relations Board.

AI’s status as a curiosity for HR departments did not last very long: The majority of C-suite leaders in a Deloitte survey published last week said they were seeking to change their talent strategies due to AI. That follows reports from other sources that stressed the need for organizations to develop AI-specific strategies for the workplace.

Littler’s survey, however, appears to show that there may still be a significant level of hesitancy in the industry around AI. Of those using generative AI for HR in the survey, the plurality said they were using the technology to craft HR-related materials or for self-service in the form of chatbots.

“Employers should be intentional about their generative AI usage, not only with regard to whether they use it at all, but also how, why and when,” Niloy Ray, shareholder at Littler, said in a press release. “Policies governing the safe usage of AI can help protect against inadvertently running up against data privacy, discrimination and other potential vulnerabilities, along with the reputational issues that can stem from AI hallucinations.”

Compliance risks in the AI space include the growing number of laws governing use of the tech to evaluate job candidates. At least three jurisdictions have such laws in place, including the states of Illinois and Maryland as well as New York City. At the federal level, the U.S. Department of Labor recently provided guidance around how AI could interact with various employment laws. Others are concerned about the potential diversity, equity and inclusion implications of AI tech.

Still, approximately 52% of all jobs are expected to change significantly due to future integration of generative AI within the next decade, according to a report by Cognizant and Oxford Economics. A February report by the Society for Human Resource Management and The Burning Glass Institute found that the tech will have the most notable effects on companies in the finance, insurance, professional services and information systems industries — with the potential for a “high human cost,” the report said.

A large majority of employers responding to Littler’s survey downplayed concerns about job displacement due to AI, however. Eighty-eight percent said they were either slightly or not at all concerned about this.

One of the biggest obstacles to expanding artificial intelligence use in business is the tricky balance between incorporating AI tools and “achieving productive coexistence” with human talent, according to a report from CompTIA, a tech industry nonprofit.

In a survey of more than 500 tech industry professionals, companies also cited issues such as higher tech infrastructure costs, cybersecurity and privacy. 

To address these challenges, companies are looking closely at skills building as part of their strategy to invest in AI-related tools for customer relationship management, business productivity and HR systems.

“These tools are already integral parts of corporate workflow, and AI will become a powerful new part of a complex solution stack,” Seth Robinson, vice president of industry research for CompTIA, said in a statement. “In this scenario, a company will likely require a low degree of AI expertise across its entire workforce and a high degree of AI expertise in a few select areas.”

CompTIA created an AI framework with six areas for skills building as AI implementation grows:

• Cybersecurity, with AI capabilities to secure networks
• Auto coding, with AI coding programs alongside human engagement
• Architecting, where humans build AI networks
• Tech-adjacent work, where staff need to understand best practices for AI prompts
• Data analytics, which will continue to evolve alongside AI
• System operations, with enhancements in business process automation

To strike the balance between AI tools and human talent, L&D can lead the transformation by putting people first, according to Peter Hirst at the MIT Sloan School of Management. Since AI tools present opportunities for innovation but also major risks, it’s important for HR to ensure organizational capabilities and competencies through AI upskilling.

However, HR teams are often left out of AI plans, despite the need for upskilling and change management, according to a McLean & Co. report. Instead, HR leaders should be at the table to help develop strategic AI implementation plans that maximize ROI and mitigate harms to reputation, security and inclusivity, the report found.

Talent leaders are also grappling with uncertainty around AI use in 2024, particularly when it comes to redesigning work to incorporate AI and automation, according to a Mercer report. The gold standard C-suites are successfully embracing AI by prioritizing human-centric productivity, fostering trust, ensuring equitable work practices and leaning into a “digitally-infused future,” Mercer said.

As companies rapidly adopt artificial intelligence tools in the workplace, HR and organizational leaders should be involved to ensure new changes can be implemented successfully, according to a report from McLean & Co. — but that’s not always the case.

“Despite playing a significant role in organizations, HR is being left out of the conversation on AI adoption,” according to the firm. That’s a mistake because HR leaders can help develop a strategic and holistic approach that considers the benefits, risks and objectives of AI use — while also maximizing return on investment and mitigating harms to reputation, security and inclusivity, it said.

“The AI strategy is the central point from which both technical and people-related AI activities originate,” Lisa Highfield, principal director of human resources technology and AI at McLean & Co., said in a statement.

“It provides the direction, guideposts, and priorities that inform the organization’s activities and enables a smooth execution to transform AI from an idea into reality,” she said. “However, introducing AI technologies can create tension within an organization, considering there are often differing views, competing priorities, and large volumes of change.”

That’s where HR and organizational leaders can play a role in aligning AI strategy with broader organizational goals, Highfield noted. A well-designed AI strategy can provide benefits such as an aligned vision and mission, proper governance, higher success rates with digital transformation and better ROI. 

Without an informed strategy, though, companies could face risks such as technological dependence, unintentional bias and threats to data security, privacy and confidentiality. 

The McLean & Co. report suggested several steps to navigate AI adoption, such as establishing an AI strategy steering committee, assessing current and target AI maturity, identifying AI use cases and values, and creating an AI roadmap based on organizational priorities.  

Importantly, the report noted, leaders should communicate AI strategy and initiatives across employee groups to keep everyone informed and involved with the plan.

Upskilling should be part of the strategic plan and communication initiatives. In fact, employees of every position and background may need AI training, experts have told HR Dive.

About 90% of HR leaders believe up to half their workforce will need to be reskilled in the next five years as AI developments roll out, according to a PeopleScout and Spotted Zebra report. To do this, leaders will need to communicate their plans, address fears and ensure employees receive the training they need, the report found.

Workers want these training opportunities, too. About 57% of workers worldwide said they’re ready for reskilling and retraining in new roles amid AI developments, according to a Boston Consulting Group report. Job seeker interest in learning and career development has increased as well.

Do you have a sense of how many people in your organization regularly use AI? Whatever you are thinking, the number is likely much higher — and the contexts might surprise you. Whether your employees use it as a research assistant, a writing aide, an image generator or a coding tool, the AI transformation of your organization is already underway.  

Good news might also be bad news  

The growing awareness and experimentation with AI tools present exciting opportunities for innovation across industries. However, alongside these advancements come new and potentially challenging risks for organizations. From an L&D perspective, ensuring that organizational capabilities and competencies around AI are robust is paramount.  

Here are three fundamental questions to consider:   

1. How can we facilitate AI upskilling and promote knowledge sharing within our organization?  
2. How do we ensure that leaders, managers and employees deploy these tools wisely, mindful of the risks involved? 
3. How can we advocate for AI education to senior leadership, recognizing that staying stagnant is not a viable option in today’s rapidly evolving technological landscape? 

Managing change is not new    

Considering the business implications and applications of any technology extends beyond understanding its technical aspects. It’s about understanding the systemic changes that it may cause or enable, and the strategies required to lead and manage those changes. L&D professionals are no strangers to managing change and this new challenge is no different than previous technology-fueled market disruptions.  

At MIT Sloan Executive Education, within many of the AI-focused courses that we’ve been developing, up to half of the curriculum is dedicated to understanding how AI impacts organizations and individuals, and the necessary adaptations in organizational structures and management practices. Addressing the human dimensions of AI’s impact is essential for fostering innovation in today’s dynamic business environment.   

But the pace of change is unprecedented    

The challenge that we are all experiencing, particularly around AI and Generative AI, is the breakneck pace of change along all the dimensions. Your list of things to consider all at once may look something like this:   

• Capabilities of the technology itself and its future potential 
• Current and emerging players leveraging these tools 
• Implications across sectors 
• Potential regulatory landscapes 
• The emergence of disruptive market entrants 

New approaches to learning design  

Traditionally, for learning programs, there’s been a very good practice of being thoughtful about designing around detailed needs analysis, documentation of intended learning outcomes, and iterative improvements. If you have that sort of traditional design model, it risks being out of date by the time any course is ready.  

Courses developed over a 9-to-12-month period can quickly become outdated before launch, necessitating constant updates. To stay ahead, we need to explore alternative approaches to learning design that foster agility. This means being able to update existing learning content continuously and swiftly develop new content as needed. 

For example, leveraging comprehensive company data analysis to craft impactful learning experiences has long been a reliable strategy. ​​But doing that methodically runs the risk of slowing you down so much that you’re not really distributing useful current content in this space. How do you solve for that? At MIT Sloan Executive Education, we’ve been experimenting with shorter education modules: more focused, but still in-depth content that satisfies the ballooning demand in nearly real-time.  

Discovering sources of expertise  

Even though we have offered AI-focused courses at MIT Sloan Executive Education for many years, the interest from organizations and individuals has mushroomed in the last 12 to 24 months. As we have been expanding our programming on the topic, we’ve been discovering that not only are there experts in AI and its applications on our faculty and in our ecosystem, but also a lot of leaders at MIT, the Sloan School faculty and researchers have been accelerating the use of these tools in their work.  

This is a hypothesis, but many organizations might find that sort of richness as well. Don’t assume that your IT function is the only expert on anything technology-related. In fact, IT could be posing constraints to the use of AI within your organization because they are legitimately concerned about the risk management and support needs.  

Filtering out the noise  

With the AI field developing so rapidly, the challenge lies in finding training programs that align with your organization’s specific business objective. As interest surges, many individuals are quickly transitioning to become voices in the AI discourse, adding to the diversity of perspectives. 

The question arises: to what extent should AI tools be leveraged to accelerate the creation and updating of learning assets? This presents both an intriguing opportunity and a significant challenge for various reasons. These technologies enable anyone to create a curriculum, allowing anyone to generate what appears to be a well-structured executive course on AI simply by engaging a generative AI tool. However, questions persist about the proprietary nature, accuracy and comprehensiveness of the content.  

How do you sort through and filter what the high-quality sources of knowledge and insight are, what the future potential is, and how fast that will come into reality? And what does employing these technologies actually mean for organizations and their workforces? It all may feel daunting, considering how fast AI technologies are evolving, but keeping your focus on the human needs and organizational needs is a good place to start.  

Employers should implement artificial intelligence technologies transparently and with the “genuine input” of workers and their representatives, the U.S. Department of Labor said in a guidance document.

DOL included these items in a set of eight “AI Principles for Developers and Employers.” The agency said the guidance would create a road map for balancing AI’s potential benefits for businesses and workers while protecting workers from potential harm.

DOL specifically recommended that AI be designed, developed and trained in a way that protects workers and that employers have “clear governance systems, procedures, human oversight, and evaluation processes” for workplace AI. The agency also cautioned that AI should not violate workers’ rights, including the right to organize, and that any data collected, used or created by AI systems be limited in scope and location.

The guidance document is a direct follow-up to President Joe Biden’s 2023 executive order directing federal agencies to develop AI “principles and best practices” specifically targeted at mitigating potential harms and maximizing benefits for workers. DOL confirmed the link between the two documents in a press release Tuesday.

“Workers must be at the heart of our nation’s approach to AI technology development and use,” Acting Secretary of Labor Julie Su said in the release. “These principles announced today reflect the Biden-Harris administration’s belief that, in addition to complying with existing laws, artificial intelligence should also enhance the quality of work and life for all workers.”

The guidance drew, at least in part, from the input of labor unions. In an email to HR Dive, the Retail, Wholesale and Department Store Union said it “contributed to the considerations” of DOL’s principles and said the guidance provides “a useful framework for understanding AI and the scope of its impact on working people and guiding solutions for reducing its impact.”

DOL said it also collected input from workers, researchers, academics, employers and developers through public listening sessions in advance of issuing the guidance.

A recent Littler Mendelson survey of U.S. employers showed that while a sizable share used generative AI for HR processes, an almost equal number had not done so. Compliance risks were a key concern for respondents, the law firm said.

While most employers in Littler’s survey said they were not concerned about the potential for AI to displace workers, other organizations — including the Society for Human Resource Management — have cautioned that AI could affect myriad workforce priorities, head counts included.

So far, regulation of AI in the employment context has come courtesy of several state and local hiring laws as well as a few federal efforts. Just weeks ago, DOL issued separate guidance detailing AI’s potential interaction with laws such as the Family and Medical Leave Act and the Fair Labor Standards Act. In 2022, the Biden administration published a blueprint for addressing bias and discrimination by automated technologies including AI.

Long before the explosive rise of ChatGPT adoption, 75% of U.S. companies were already using some form of artificial intelligence in the employment lifecycle. The use of large language models as a component of the employment decision-making tool kit is a sophisticated, albeit less transparent, step forward that extends beyond traditional resume screening.

In emerging iterations, the technology permeates nearly every stage of the employment process, from hiring and compensation to promotion and termination. Regulation has responded accordingly, increasing at local, state, national and global levels. In this environment of technological and regulatory change, it’s imperative that companies examine how AI tools, particularly opaque systems with hidden layers, interact with employment data to mitigate risks and promote beneficial outcomes.

Many organizations are not aware of the extent and nature of AI use across their business. This can be problematic, as algorithms are only as reliable as their training and integrated data allow. Whether intentional or inadvertent, the biases of software engineers can (and do) find their way into software coding, leading to negative outcomes (e.g., employee homogeneity). Regardless of who developed or sold a program, or whether a company is aware of how a program does what it does, it is the company employing AI that will be held responsible for any misuse.

Real world risks

The harm caused by AI is no longer a future concern. It’s a reality today.

A legacy hiring test established by the EEOC, known as the “four-fifths rule,” also referred to as the Adverse Impact Ratio, continues to set the standard for how to determine adverse impact in hiring practices.

By calculating the difference in hiring rates between favored (highest selection rate) and disfavored groups, this rule can also be applied to review discrimination within AI. For example, if a job application input is for the candidate to be able to lift a certain amount of weight, AI review of that application could blindly rule out candidates of certain age or gender. In this example, the four-fifths rule would surface any disparities in hiring rates across the age or gender.

Similarly, the EEOC has flagged concerns over AI automatically rejecting disabled candidates or individuals with employment gaps. More recently, the agency issued guidance regarding employer use of AI in any aspect of the employer selection process, including hiring, promotion and termination.

Though AI is mathematically sound, when fed empirical inputs, it may become overly reliant (i.e., “overfit”) on the data on which it was trained, not allowing accommodation of new data, and in turn producing homogenous results (e.g., a biased candidate pool). To avoid this, organizations must ensure their training data is sufficiently diverse to allow the model to identify a diverse hiring pool. If a company is using a black box tool, the system could begin incorporating its own decision making, compounding the overfit issue without anyone knowing this is happening.

As another example, a recruiting team might believe they have eliminated all possible points of discrimination when it comes to a subject such as race or other overt characteristics that may lead to discrimination. However, if a factor such as zip code is introduced, and there is a correlation, race may be encoded by proxy. There can also be several hidden layers in an AI system, where the software has a foundational knowledge base that is known. Yet, unknown to the user, it could continue creating its own variables and seeking out real-time data from other sources to achieve its primary goal of providing an outcome.

There are clear regulatory risks associated with discriminatory and unfair employment practices. Moreover, a business can also be damaged over the long term through a decline in diversity, lack of visibility into a wider talent pool, diminished company culture and public scrutiny for AI governance failures.

Mitigating actions

An AI system must be designed and implemented to avoid bias, treating individuals equitably, with as much transparency as possible. Governance should leverage existing structures and be supported by a committee that can guide AI usage within risk management frameworks. Upon that foundation, there are several steps organizations can take to help realize the benefits of AI and mitigate misuse:

• Craft a purpose statement. What is the company working to achieve at the highest level?
• Define organizational objectives that support the overarching goal.
• Develop an AI program framework or policy and rules-based metrics. Ask questions like, “what does this mean for the organization over time?”
• Consider outcomes in a meaningful way, i.e., whether approved use cases are aligned with the risk portfolio and purposes.
• Adapt the framework as needed to remain aligned with company purpose. There must be ongoing human oversight, i.e., whether approved use cases are implemented and maintained in accordance with approved governance structures.
• Address major changes in company structure and data inputs that may impact AI models and governance frameworks (such as during M&A activity).

Legal, compliance and HR teams should be prepared to provide a clear explanation of how AI software is being used and confirmation that it is aligned to regulatory requirements and data protection everywhere the company operates. Proactive and consistent testing must be upheld to continuously assess and address instances of bias.

The regulatory environment will continue to evolve. The EU AI Act is only the beginning. President Biden issued an executive order on AI that delegates action to certain agencies, with different regulators tasked with various aspects of assessment and enforcement. New York City launched its own AI action plan in late 2023 and several U.S. states have introduced legislation. More will follow.

A 2023 Gartner survey of HR leaders said more than 60% are engaged in enterprise-wide discussions around the use of generative AI.

More than half are doing so in collaboration with IT leaders, yet fewer than half (45%) are working with legal and compliance teams. It’s critical for legal and compliance to be involved in these decisions early and throughout the process of implementing new tools to ensure it is done responsibly and in compliance with company policy and regulatory guidance. Through strong governance and the continued oversight of human discernment, AI can be a powerful tool for HR teams and other functions within the business.

Employers can boost motivation and learning by giving their employees the autonomy to overrule artificial intelligence decisions from the start, according to a recent study by researchers at the University of Auckland and the University of Western Australia.

In addition, combining early autonomy with incentive pay can create an environment where employees feel more engaged and learn faster, they found.

“Imagine you’re a financial specialist at a bank. You input details for a mortgage application, and the AI system recommends declining it,” Frank Ma, one of the study authors and a lecturer at the University of Auckland Business School, said in a statement. “While the system is based on hard data, as a human, you can recognize nuances — ‘soft’ information — that AI can miss. This is where the ability to overrule the system is beneficial.”

In the study, 161 participants completed a series of online tasks, which were designed to simulate real-world scenarios in sectors where AI is used for managerial decision-making.

The researchers found that when participants felt empowered to overrule AI decisions early, they became more motivated and better equipped to handle complex tasks. Giving immediate flexibility allowed for more opportunities to override system decisions, and combined with incentive pay, participants were more likely to put in more effort to make the final call accurately.

On the other hand, delaying the autonomy to override AI appeared to have a detrimental effect on employee motivation, which could limit learning and performance.

“Employees with incentive schemes and immediate flexibility get a better understanding of their roles and improve their performance,” Ma said. “We believe this is due to developing a more in-depth understanding of how the computer system or AI generates its decision.”

The U.S. Department of Labor has recommended that employers include workers in their AI plans and implement the tools with transparency. The department said AI should be developed and trained to protect workers and that employers should have “clear governance systems, procedures, human oversight, and evaluation processes” for workplace AI.

At the same time, most business leaders believe they can’t train employees quickly enough to keep up with AI and other tech developments during the next three years, according to a World Employment Confederation survey. Most leaders also said these tech disruptions will require companies to rethink skills, resources and new ways of doing work to cover the technical and soft skills gaps.

However, HR teams are often left out of AI plans despite the need for upskilling and people management, according to a McLean & Co. report. HR leaders could help develop a strategic approach that maximizes return on investment and mitigates harms to reputation, security and inclusivity, the firm said.