Some observers predict technology that eliminates career, leading to mass redundancies and unemployment. Supporters, on the other hand, predict that AI will lead to a golden age of work, releasing time for workers to spend more complex and surprises.
The actual answer will not be found in headlines, but in research, he says Hatim A. RahmanAssociate Professor of Administration and Organizations in Kellogg. “We need to understand what trends we are recreate and what trends we want to change.”
In a new review document, Rahman and his colleagues analyzed more than 300 articles of magazines to tackle the cumbersome but urgent issue of AI inequality and workplace. By connecting decades of working in all sectors, they created a framework for today’s researchers and scholars to better understand how AI activates and enhances the inequality in the workplace as a waterfall.
Transfer reflects how seemingly small decisions at different stages of the AI life cycle can snow quickly. The options made during the design, application or use and adaptation of the model can range to aggravate or reduce the workplace inequality.
Using this context, researchers can identify the critical points where it would be more effective to intervene and minimize AI damage or to enhance its benefits, the authors write.
“The results of AI are going to play out for several years, if not decades, because it is complicated and will vary quite a bit,” Rahman says. “We thought it was really important to be able to associate how technology develops and applies to these different forms of inequality.
One target, four perspectives
As Rahman and Coauthors Arvind Karunakaran and Devesh Narayanan at Stanford University and Sarah Lebovitz of the University of Virginia attended previous literature on how AI and other technologies affect the workplace inequality, they found that it has attracted researchers. As soon as an area dominated by computer science, AI Research is now attracting experts from economics, psychology, sociology, organization and management studies, philosophy, law and politics. The authors organized this wide range of views in four separate perspectives, each focused on different consequences from the introduction of AI to a workplace:
While these perspectives are examining inequality using different expertise, their conclusions often overlap, the authors have found. The establishment that common ground can prevent repeated historical errors.
“For such an important issue, we can’t afford so much,” says Rahman. “In the past, economists may not have spoken to sociologists and vice versa and sometimes revealed the same ideas using different methods and terms. Thus, one purpose of this review is to truly try to encourage interdisciplinary discussion and bridging.
Phenomena
Instead of treating these four types of inequality as AI’s separate potential results, Rahman and your partners suggest that they connect them in succession to the transport of a waterfall. This frame reflects that different points along the AI life cycle can cause a series of events that can climb out of control if left uncontrollably.
For example, decisions made during the design of an AI model may have unpredictable consequences. A system for controlling work candidates trained in biographies by a population that cannot lead to discrimination practices. An evaluation system built to evaluate the performance of the work of white collar workers can score harder workers with blue collar, creating wider pay gaps.
Instead of looking at exactly what a AI model is doing, Rahman says, “We must ask,” Who designed it? What were their values and priorities when designing the model? “Design and development, at least in my field, is the one area that is overlooked most and very rarely addressed.”
The application is another area where the seeds of inequality in the workplace can be planted. Failure to engage in important stakeholders in this process can reduce confidence in AI or enhance structural or relational imbalances. Rahman reports the ongoing reaction against Chatgpt by artists and teachers who felt that their voices were not heard in the early stages of technology.
“There is nothing a priori that suggests that chatgpt and genetics had to be released this way,” Rahman says. “Why does the copyright law not adhere to? [for AI]; ”
As a result, many artists and teachers perceive AI as a tool for their replacement, instead of a tool they can use to improve their work. This skepticism can deepen the inequalities between management and labor, high wages and low wages and other teams in the workplace.
“After all, many of the decisions on how a technology is to be used or AI are not related to the capabilities of technology,” Rahman says. “[They are] Show much more the ideologies, the value we place at work and who is in the room when these decisions are made. ”
An upstream research agenda
Through the AI explosion, there are countless questions about the inequality in the workplace that researchers can choose to follow. What kind of jobs are most affected when technology is applied? Does it offer lower ranking employees the opportunity to “upgrade” their jobs by taking more complex tasks? What factors inspire workers to experiment with AI instead of reacting with fear and concern? What new jobs do you create?
The concept of AI inequality cataract can help direct attention to the most aggressive questions, Rahman says. By focusing on the parts of AI’s life cycle where inequalities arise, researchers can help workers and policy -executives to understand and prevent these consequences.
But to make the biggest difference, he suggests look up, in the first steps of design and implementation. This is easier to say than to be done when commercial technologies are based on narrow kept secrets and randomized controls are unlikely.
“There is so much more device for studying the cathedral,” Rahman says. “But inequality is not a simple thing to measure or tease outside, so we need sophisticated theory and data collection around it.”
Inequality waterfall can also slow down or minimize with its environment with a circle. Applying each of the four prospects for research on application and adaptation phases can help to determine in detail how and why inequalities occur. Experts can then use these ideas to update the design of future AI systems, preventing the same consequences from the repetitive and spiral.
And while AI seems to be moving in the workplace at a speed that does not allow careful examination, Rahman says there is still time for researchers and business leaders to do it right.
‘Some of the recent estimates suggest very moderate profits over the next 10 years About AI, “Rahman says.” I think it shows a lot of opportunities and challenges for researchers to really understand and tease the AI relationship and result in the workplace. “
