Artificial Intelligence (AI) has been called the most transformative technology since the invention of the digital computer. The “AI revolution” has even been compared to the agricultural revolution and industrial revolution in its potential to change the way people work.
While the full impact of AI on people’s lives is impossible to predict, its effect in the workplace can be seen in many subtle ways – primarily by altering or even reinventing processes across industries and throughout the entire value chain.
Impact of AI on the contingent workforce
The effects of AI in extended workforce management are just beginning and can be seen shaping how jobs are described, how candidates are evaluated, and how contracts are written. AI will soon be employed for more strategic outcomes like workforce planning, workforce optimization, and total workforce management.
When AI-assisted tools are used to source or manage contingent workers, the value they deliver tends to fall in three areas:
- Lower costs
- Better quality
- Increased efficiency
Cost reduction
AI models can be used to replace survey-based outdated rate cards with market-relevant rate cards by analyzing large data sets of market rates based on skills, job titles, and geographical locations. AI can uncover overspending and compliance issues related to misclassification. It can even be used to predict and guide novice buyers to negotiate better bill rates and fill assignments faster.
Higher quality
AI-powered tools can evaluate contractor, supplier, and program performance. They can set benchmarks and inform managers who should be redeployed, and which service providers deliver the best value in terms of both budget and schedule.
Increased efficiency
AI is particularly suited to accelerate time-consuming manual processes. From resumé review and screening (80% time reduction) to job taxonomy setup and refresh (85% time reduction), AI dramatically cuts administrative tasks, allowing HR and procurement professionals to focus on responsibilities that add more value to the organization.
“We have an ecosystem of 5 million workforce candidates. It would take us 7 days to identify candidates. With Beeline, we can do that in an hour.”
- Greg Muccio, Southwest Airlines
What to look for in evaluating AI applications
AI can deliver significant benefits in the sourcing and management of external talent. But it also poses certain risks. In addition to legal compliance risks, there are ethical and operational risks that can affect your contingent workforce program and your company’s bottom line and reputation.
To minimize risk in the use of AI-powered applications, it is important to keep three principles in mind:
- Human in the loop
- Ethical and compliant
- Trustworthy
Human in the loop
It has frequently been said that sustainable value doesn’t come from technology itself, but from what people can accomplish with it. We work in a people-centric business, so it is important for people to direct and manage AI-based applications, and not let AI direct itself.
Over the years, AI-powered applications have been shown to reflect bias, develop bias, and even amplify bias. The US Equal Employment Opportunity Commission, which has focused on AI’s role in hiring since 2016, investigated a complaint that a digital recruiting company using AI for facial recognition during candidate assessments was unfairly biased. As a result, the company terminated the facial analysis component of its screening assessments. In 2021, New York City passed a law (which went into effect in 2023) requiring employers using AI-based candidate screening tools to undergo an independent audit to check for bias against protected groups.
To guard against algorithmic bias and unintentional but potentially harmful results, users must ensure that AI supplements, but does not replace human decision-making. Human control must be built into both the AI-enhanced applications used and the processes they empower.
Ethical and compliant
AI-assisted programs must comply with the growing body of laws and regulations, and with companies’ codes of ethics. They must also respect the needs, rights, and values of the individuals whose data is entered into the contingent workforce program.
It is not enough just to comply fully with legal requirements. AI-enhanced applications must also be designed and monitored to ensure that they respect human dignity, privacy, human rights, and civil liberties.
Trustworthy
AI methods, applications, and uses must be trustworthy. To be worthy of trust, they must be transparent and accountable. A “black box” approach to AI is unacceptable.
Transparency requires that the program is understandable by a human being. It must be possible to determine how the AI model is built, how it works, and how it mitigates against errors and all forms of bias.
Data transparency is also required for the process to be trustworthy. The value and validity of AI depend upon its data sets. Data is the lifeblood that fuels AI algorithms, allowing them to learn, adapt, and make decisions.
A major risk for users of AI-powered applications is the size and quality of the data sets that drive them. Bad data – wrong data or too small a data set – can result in bad decisions and invalidate the use of AI altogether. That’s why a transparent data set is critical to a trustworthy outcome.
