The world’s top corporations are either building their own artificial intelligence (AI) tools, or rushing to adopt existing solutions to optimize their operations. But at what cost?
Google knows the cost of a mistake – the tech giant saw its shares tumble 9% after its newly released AI chatbot Bard gave inaccurate answers in a promotional video.
But in the long run, it is people who will pay the price for the rash uptake of AI tools.
The accounting firm, PricewaterhouseCoopers (PwC), signed a 12-month contract with AI startup Harvey to “speed up” the work of its 4,000 legal professionals.
Harvey’s AI solution, which relies on OpenAI’s technology, will not provide legal advice or replace lawyers, Reuters reported the firm insisting.
Well, maybe not for now. However, experts and human-rights watchdogs are already sounding alarm bells.
This week, STAT news outlet, which focuses on health, medicine, and life sciences, ran a story on how the Medicare Advantage insurance firm relies on AI, cutting off care for patients who can’t even dress themselves or go to the bathroom.
Dangers of human analytics
“Employees must be included in the implementation process, if so-called artificial intelligence systems are introduced to their workplace,” AlgorithmWatch, a non-profit research and advocacy organization, insists.
The group, closely monitoring automated decision-making (ADM) systems and their impact on society, released a study on AI in the workplace, where experts urge employers to give workers more control over automated systems to prevent their being marginalized.
“Automated decision-making systems in workforce management undermine established processes that ensure worker participation. Without an insight into these systems, employees exposed to them more often than not remain powerless. For this reason, comprehensive transparency has to be introduced as a standard. Furthermore, employees must be included and have a say in every process concerning their workplace,” study co-author Dr Anne Mollen said.
The study looks at the effect increased implementation of ADM can have on talent and the labor market. Its authors refer to such systems as People Analytics, also known as Resource Analytics and Workforce Analytics – all terms describing software-based automated systems that manage a company's workforce.
“Common fears concern job loss and de-skilling, due to increasing workplace automation, the dangers of discrimination and false decisions made by ADM systems, or fears about the increase in workplace surveillance and violations of privacy,” the study reads.
Its authors rely on the term ADM instead of AI to “avoid the pitfalls created by mystifying the [latter] term”.
The study says ADM systems are already being used to scan CVs during hiring, allocate shifts, automate performance reviews, and even decide whom to lay off.
Researchers argue that the further adoption of these systems can create an even bigger imbalance in the workplace, since employees are increasingly subjected to ADM and cannot evaluate whether its verdicts are fair and based on proper data sets.
In the study, AlgorithmWatch proposes a model for involving employees in People Analytics systems, outlined in five steps.
“ADM systems used for workforce management cannot be made to work in the interests of employees without ensuring the continuous involvement of employee representatives throughout the entire development, implementation and application phases of these systems,” they said.
More from Cybernews:
Subscribe to our newsletter