Robots are taking over our world. Many of us live with this fear, unaware that the words augmentation and automation are indeed describing a different level of human interaction with machines. So, automatically, the question comes up: Will humans soon be redundant?
Robots are better at everything. Apparently, as they are taking jobs away. Hence, they are better drivers, they register our groceries in the supermarket, and they collect and analyse our data (Thomas, 2020). Radiologists are made redundant, so are document review lawyers, as well as cashiers. Soon we will be left with employees with zero economic value (Autor, 2015). Watson will educate our children, while Google will drive us around, and we will soon be redundant as robots will live our lives for us (Colvin, 2014). These statements probably sound familiar.
Before analysing the need and work of robots and their take-over, we should differentiate between automation and augmentation and define each term. Based on Merriam-Webster, automation is the “automatically controlled operation of an apparatus, process, or system by mechanical or electronic devices that take the place of human labor.” The same source defines augmenting as the act “to make greater, more numerous, larger, or more intense.” Thus, one act replaces the human, and the latter enhance human abilities and is considered an addition to the human brain (Davenport & Kirby, 2016).
Davenport and Kirby (2016) illustrate the advantage of automation and augmentation by demonstrating the self-checkout counters at grocery stores. In old times, cashiers had to memorise prices for every product and errors were traced to the human brain’s weakness. Now through automation, self-checkout counters ensure accurate prices for goods and promise shorter lines. Previously, five cashiers had to work while now only one person is in charge as a supervisor. However, Davenport and Kirby (2016) state further that a previous augmentation through the introduction of scanners has enhanced the cashiers´ work by assisting them in their weakness. Thus, while one could argue that the automation eliminates the cashiers´ work, it fails to consider that customers do not always use the self-checkout and would still rely on humans, even though the error propensity of the human brain is known. Hence, the scanner is only a small change to our previous cashier experience, which Iansiti (2020) calls “weak AI”. It reshapes the general supermarket business model by ensuring growth in the correct digital direction.
Another example would be “Blockbuster”. The former innovation of family video stores, which were designed to have a friendly atmosphere, was originally designed to be different from the dark back rooms of Triple X video stores. However, when Netflix entered the market in 2000 (Austin, 2016), the business model for video rentals changed. Customers were given the convenience to stream movies directly, without leaving their homes. Movie streaming was automated and this disruptive innovation created a new business model, eliminated Blockbuster, and subsequently made the employees redundant. But again, using this statement as is fails to evaluate whether Blockbuster’s bankruptcy was based on automation or rather on the fact that it failed to see the digital change.
Another area that was affected by automation and augmentation is legal technology and eDiscovery. EDiscovery has started the world of legal technology. Twenty years ago, pre-discovery evidence had to rely on human work (Austin, 2016). The human document reviewer managed the market with their language skills, understood the case, and subsequently determined the hourly compensation. In 2008, when I started working as a document reviewer, the general price for an English-speaking reviewer was USD 75 per hour. Each additional language would add USD 50 to this hourly rate.
However, through the application of Artificial Intelligence (AI), computers started reviewing documents in an effective and efficient manner. This led to the elimination of unnecessary human employees and decreased any overhead. Once the documents underwent AI, the results were reviewed by attorneys. The hourly price for an English-speaking review lawyer has decreased to USD 20 per hour. Therefore, AI has not caused automation, but an augmentation, as the computers became co-workers of the human reviewer. As Davenport (2016) states, humans created these machines and will use them to ease their work, not to make humans redundant. Thus, the last step in any completion process will always remain human decision making. AI in predictive coding will ease lawyers´ work to ensure that they focus more on the legal aspects or analyse the data concerning the case context. However, this will not eliminate the human work or lead to shorter workweeks. It will solely shift the focus of the human work, as the machine will never be able to learn without the human input. Automation of the Electronic Discovery Reference Model (EDRM), primarily of the review aspect, would require computers to collect, process, and review the documents without any sort of human input. This could be comparable to the self-check counter at the grocery store. Humans and computers work hand-in-hand and each one is leveraged for their strength (Autor, 2003). This requires that the human receives sufficient and continuous training. The automation of specific workflows requires the identification of job skills and job fits (World Economic Forum, 2018).
Similar to the machine that strives to get closer to Six Sigma, the human worker needs to upskill in order to stay above the machines (World Economic Forum, 2018).
Learning
The development and continuation of the new industrial revolution through automation as well as augmentation have changed the working environment. But this change enforces us to consider the opportunities it has brought. Jobs have not disappeared but required working skills, and requirements have adapted and adjusted to the new needs of the market (World Economic Forum, 2018).
Some twenty years ago we were referring to office assistants, review lawyers, and supermarket cashiers, just to give a few examples. These employees had to do basic work which can now be done by machines. Thus, only the basic opportunities are taken by the automation process. However, adding skills and endless learning opportunities are enabling human workers to grow into different branches of their original work (Dickie, 2020). Litigation Support, eDiscovery Specialist, and Processing Analyst are some of the positions that did not exist before humans and machines started working hand in hand.
Table 1
But it should be mentioned that job seekers and present employees must differentiate themselves by stepping up and evolving from the basic worker (Dickie, 2020). From the stepping role table (Table 1) below, review lawyers can see the development of their job role and present requirements. While 2008 required legal and language skills, now the Litigation Support Analyst or the Document Review Manager has to have technical, management, and people skills. With the newly gained knowledge and the additional skills, human workers are getting more in demand (Autor et al., 2003).
Conclusion
Automation can only lead to the right results if it does not require a subsequent consumer-human interaction, but an interaction of human and machine during the entire business relationship. Thus, work that is focusing on emotional intelligence, needs quality control, has to analyse a big picture, and is overall responsible for the design of a machine, cannot be automated (Chui et al, 2016). As soon as the consumer has the option to type “ zero” in order to speak to a human customer service assistant or speak to a human Document Review Manager, the complete automation is interrupted and it is an augmented workflow. Thus, in the latter case, humans and computers complement each other (Autor et al., 2003; Fourie, 2016). As Davenport (2016) states, the former chairman of the US Federal Reserve was denied a loan application based on his financial input. The machine did nothing wrong by denying the application based on the numbers. However, the machine was unable to identify that the former chairman’s finances did not follow a straightforward financial pattern. Only a human could have made the distinction.
Davenport and Eberly (2016, p.63) state “The general point to be made is that machines need augmenting when there are important exceptions to rules and structured logic.” Autor (2003, p.25) adds to this statement by suggesting “Tasks that cannot be substituted by computerisation are generally complemented by it. This point is as fundamental as it is overlooked.”
About the Author
Dr. Nina Mohadjer, LL.M. has worked in various jurisdictions where her cross-border experience as well as her multilingual capabilities have helped her with managing reviews. She is a member of the Global Advisory Board of the 2030 UN Agenda as an Honorary Advisor and Thematic Expert for Sustainable Development Goal 5 (Gender Equality) and the co-founder of Women in eDiscovery Germany.
References
- Austin, D. (2016). Evolution of eDiscovery Automation, AECD webinar, https://www.aceds.org/page/certification
- Autor, D. H. (2w015). Why Are There Still So Many Jobs? The History and Future of Workplace Automation†. Journal of Economic Perspectives, 29(3), 3–30. https://doi.org/10.1257/jep.29.3.3
- Autor, D. H., Levy, F., & Murnane, R. J. (2003). The Skill Content of Recent Technological Change: An Empirical Exploration. Quarterly Journal of Economics, 118(4), 1279–1333. https://doi.org/10.1162/003355303322552801
- Chui, M., Manyika, J., & Miremadi, M. (2016). Where Machines Could Replace Humans–and Where They Can´t (yet). McKinsey Quarterly, 3, 58–69.
- Colvin, G. (2014). In the Future, Will There Be Any Work Left for People to Do? Fortune, 169(8), 193–202.
- Davenport, T. & Kirby, Julia (2016). Only Humans Need Apply, Harper Collins.
- Dickie, J. (2020). Is Your Business Tapping into All Available Data Sources? Acquisition, Automation, and Augmentation Can Help Overcome Data Management Challenges. CRM Magazine, 24(3),4.
- Fourie, J. (2016). Automation to Fuel Unemployment? Finweek, 4.
- Iansiti, M., & Lakhani, K. R. (2020). Competing in the Age of AI. Harvard Business Review, 98(1), 60–67.
https://www.merriam-webster.com/dictionary/automation (Accessed December 13, 2022)
https://www.merriam-webster.com/dictionary/augmentation (Accessed December 13, 2022) - Strategies for the New Economy Skills, 2018. World Economic Forum.
- Thomas, Z. (2020). Coronavirus: Will Covid-19 Speed up the Use of Robots to Replace Human Workers? https://www.bbc.com/news/technology-52340651WEF,2019
