The Ethics and Economics of AI
Artificial Intelligence (AI) and Machine Learning (ML) are catchy buzzwords these days. From tech gurus to software engineers from businesspeople to professors, many are advocating for the use of machine learning and artificial intelligence in a variety of fields. But what exactly are artificial intelligence and machine learning? Are there drawbacks? How can we best employ AI and ML to meet our needs? Artificial intelligence is traditionally defined as, “…the ability of a digital computer or computer-controlled robot to perform tasks commonly associated with intelligent beings.” (Artificial Intelligence). Machine Learning is a subset of the AI field. “You can think of an ML algorithm like a recipe, where the ingredients are data, and the dish is a model that can make predictions or inferences about the future.” (“What Is a Machine Learning Algorithm?”). In practice, this has led to AI/ML being developed today using large datasets to teach models how to predict things. Early on, models will often incorrectly predict outcomes, but will then update themselves with the new knowledge until they are able to more correctly predict outcomes. There are potential drawbacks with AI as it only looks at data and does not have the intangibles that humans use in their judgment. Another drawback of AI/ML that many critics often bring up is the idea that increasingly accurate AI/ML models will eliminate human jobs. This is a fear that is rooted deeply in society, causing backlash to many new technological advances. It also raises the ethical issue of people vs. profits. Many fear that companies will lay people off in favor of a better bottom line. Machine Learning and Artificial Intelligence models can be used across a broad range of business applications in nearly every industry that give them a serious use case without eliminating jobs, but could suffer from drawbacks as the technology is being worked on, including rigidity in their predictions.
Uses and Drawbacks
One example of possible drawbacks with AI/ML is algorithmic bias. Algorithmic bias is defined as, “Algorithmic bias, in the simplest terms, is a systematic error in predictive computation. In some contexts, the term bias describes statistical mistakes that predictive models make because of code bugs, poor model selection, inappropriate optimization metrics, or suppressed data.” (“What Is Algorithmic Bias?”). Algorithmic bias is one of the most important drawbacks to consider when employing AI/ML algorithms for business purposes. A use case for AI/ML in business is for inventory management. With proper modeling, AI/ML algorithms could help to predict inventory needs for a business far in advance, allowing the business to manage its inventory in a timely and efficient manner. This would limit delays and shortages of materials with advanced predictions. (The Economics of Artificial Intelligence). However, if the algorithm has bad data fed to it, this could cost the business a fortune if they overstock or understock their inventory purely based on the model. While the model is being perfected by constantly being fed data, it would be important for a human to continue to oversee inventory, providing a second set of eyes for everything the algorithm is suggesting before it is blindly signed off on.
Ethical Concerns with AI/ML
Augmenting skills through the use of AI/ML could lead to a huge change in the workplace. This idea, called Augmented Intelligence or Collaborative Intelligence, can trace its roots back as early as 1997, when chess grandmaster Garry Kasparov partnered with a computer to take on an opponent in a chess match. The key takeaway from this experiment and other, similar cases were illustrated by the Harvard Business Review, “Like with human teams, the power of working with an AI comes from how the person and computer complement each other; the best players and most powerful AIs partnering up don’t necessarily produce the best results… The tournament was won by a pair of amateur American chess players using three computers. It was their ability to coordinate and coach effectively their computers that defeated…” Collaborating with AI/ML software and models could lead to higher productivity and the automation of routine work for many. Although AI/ML will have many disruptive effects early on (as is common for major technological advances), those who are able to find and create value out of this technology will be far better off in the years to come. (Cremer and Kasparov). This technology is not going away.
So what is AI not capable of, yet? “ AI isn’t capable of developing complex strategies or thinking critically through complicated scenarios. There is a certain element of human intuition that’s critical, and many people will turn to AI to assist them in thinking through problems — but ultimately, humans will make the decision. “ This further solidifies the idea that humans will take on a critical role in interpreting AI/ML results and making decisions based on the outputs of these models. The current AI/ML models we have are still too rigid to replace the soft skills that humans have developed. This leaves the door open to creating many jobs that will help to improve these models, a key sector for job creation in the wake of the jobs that many AI/ML models will eliminate over the next few years. For example, positions for machine learning engineers, robotics engineers, and data scientists have already grown in demand over the last several years. It is also creating positions for AI trainers, which will help to further refine these models and enhance their capabilities through human interpretation. Another new sector of jobs will be AI maintenance people, who will be needed in droves in order to maintain AI/ML infrastructure. (Ascott). The massive need for upskilling across many industries to create workers capable of handling and collaborating with AI/ML will likely fall to employers and organizations. Putting employees through on-the-job training and re-skilling with these technologies will be an up-front cost for many organizations. However, this investment will pay off with the added productivity and efficiency that AI/ML can offer to workplaces. It is also a worthwhile investment for employees as they will enhance their career development and make them more capable and qualified for future career advancement opportunities. Training on AI/ML by companies for employees creates good value for both sides and will serve to boost both employee retention and morale in the face of the uncertain situation that AI/ML may create for many employees.
In fact, companies that deploy AI solely to automate processes and eliminate current employees gain only short-term productivity benefits. Firms that combine AI with current human workers see the most significant benefits. Harvard Business Review again elaborates on this concept, “Through such collaborative intelligence, humans and AI actively enhance each other’s complementary strengths: the leadership, teamwork, creativity, and social skills of the former, and the speed, scalability, and quantitative capabilities of the latter. What comes naturally to people… can be tricky for machines, and what’s straightforward for machines… remains virtually impossible for humans. Business requires both kinds of capabilities.” AI/ML serves to boost analytic and decision-making capabilities for humans, but can also improve creativity with its ability to generate countless options for the user. (“Collaborative Intelligence: Humans and AI Are Joining Forces.”). This kind of use has already been seen broadly with the use of Open AI’s ChatGPT, which can assist people in coming up with ideas for a paper, a design, a pitch, and countless other projects. Although it cannot often create a final product by itself, ChatGPT can help to spur ideas for users. This is yet another case where we see the collaboration between AI/ML and humans creating the most value as opposed to fully automated processes or fully human processes.
Combining AI/ML and Human Judgment to create value
The combination of human judgment and AI/ML algorithms data and predictions can be a powerful tool, particularly for businesses. However, without perfect data (which is a rare find) and perfect models, it will be a long time before we can rely solely on AI/ML to help run businesses by themselves, if ever. For now, it is up to humans to interpret AI/ML outputs and use them as support for decision-making rather than the sole reason for business decisions. Humans better augmenting their skills to adapt to AI/ML in the workplace will help to create more value for businesses, particularly as AI/ML models are still being trained to create reliable outputs. A symbiotic relationship between humans and AI/ML in the workplace is the best way forward in our current environment to create value and ensure that jobs are secure for the foreseeable future.
Citations:
Artificial Intelligence | Definition, Examples, Types, Applications, Companies, & Facts | Britannica. 22 Feb. 2023, https://www.britannica.com/technology/artificial-intelligence.
Ascott, Emma. “AI Will Create 97 Million Jobs, But Workers Don’t Have the Skills Required (Yet).” Allwork.Space, 19 Nov. 2021, https://allwork.space/2021/11/ai-will-create-97-million-jobs-but-workers-dont-have-the-skills-required-yet/.
“Collaborative Intelligence: Humans and AI Are Joining Forces.” Harvard Business Review, 1 July 2018. hbr.org, https://hbr.org/2018/07/collaborative-intelligence-humans-and-ai-are-joining-forces.
Cremer, David De, and Garry Kasparov. “AI Should Augment Human Intelligence, Not Replace It.” Harvard Business Review, 18 Mar. 2021. hbr.org, https://hbr.org/2021/03/ai-should-augment-human-intelligence-not-replace-it.
Robots and AI Taking Over Jobs: What to Know | Built In. https://builtin.com/artificial-intelligence/ai-replacing-jobs-creating-jobs. Accessed 13 Mar. 2023.
Simon, Charles. “Council Post: As AI Advances, Will Human Workers Disappear?” Forbes, https://www.forbes.com/sites/forbestechcouncil/2022/06/28/as-ai-advances-will-human-workers-disappear/. Accessed 13 Mar. 2023.
The Economics of Artificial Intelligence | McKinsey. https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-economics-of-artificial-intelligence. Accessed 13 Mar. 2023.
“What Is a Machine Learning Algorithm?” Center for Critical Race + Digital Studies, https://www.criticalracedigitalstudies.com/peoples-guide-posts/i4koqeh07dl5662rzovkttxg63wtw9. Accessed 13 Mar. 2023.
“What Is Algorithmic Bias?” Center for Critical Race + Digital Studies, https://www.criticalracedigitalstudies.com/peoples-guide-posts/what-is-algorithmic-bias. Accessed 13 Mar. 2023.
