An Introduction to AI in Business and Investments
It’s a bright, sunny morning in June 2027, and you’re just waking up. You check the clock, it’s 8:30am. Wow, you just slept right through your alarm. You quickly stagger out of bed and run down the stairs to the kitchen because you have an interview for a new job in less than an hour. Your coffee just finished brewing thanks to your smart home, of course. You’re starving, and check the refrigerator to find a yogurt and granola prepared precisely in your desired proportions. Fortunately, your refrigerator knew a few days ago that you were running out of breakfast food so it placed a small order online. As you sit down to eat, however, you suddenly lose grip of your coffee mug and the coffee spills all over the floor. You curse, but realize it’s ok since you probably didn’t have time to finish it anyway. Your portable robot cleaner detects the spill, and cleans up the mess. You rush back upstairs, get dressed, and run out the door. You forget to turn off the lights and lock the door, but it doesn’t really matter because the minute you closed your door, your house locked the door behind you, pulled the blinds, turned off the lights, reduced the heat, and even turned off your coffee maker. You’re an absolute mess, but thankfully your home technologies saved you on this important day.
When you think about artificial intelligence, perhaps something similar to the story above first comes to mind, or maybe something completely different. Maybe you think of an advanced machine, a sophisticated software, or some crazy robot instead. This is because AI is a concept still quite new to us. Despite the fact that AI has essentially become the latest tech buzzword spanning from the Silicon Valley all the way to China, it is a term more often spoken than thoroughly understood. Regardless, we can all agree that AI has already begun to revolutionize various aspects of our life, and the discussion of AI elicits a broad range of emotions. When I think of AI, I am typically filled with ambivalence; while I am excited about the endless opportunities that are made available through machine augmentation, the overwhelming uncertainties regarding the implementation of AI can be discomforting. This is why I hope to shed some light on the concept of AI in general, as well as unravel some of its intricacies with a focus on what the implications are specifically for the business world.
To understand AI, it’s important to understand what it is not. Many investment firms today utilize various software programs to their advantage in order to analyze data and perform various data analysis. This software enables companies to create portfolios of the “best” stocks that meet criteria related to corporate results. However, these types of software are not AI due to the fact that they are considered to be static; in other words, they repeat their function unless a person makes an adjustment. AI, on the other hand, is all about machine learning, or when a program makes adaptations completely on its own based on previous data and/or experiences. Essentially, the overarching goal of the program remains consistent, but the problem-solving capabilities continuously change in order to identify the most successful and practical solution.
AI is already changing the ways in which we go about completing various tasks in the business environment. This is why it is critical for businesses to at least understand what AI capabilities exist right now. AI is set to significantly contribute to the global economy in the future through increased efficiency and scale, so in order for businesses to avoid falling behind their competitors, it is vital for them to keep up with the fast-paced technological growth we are currently experiencing. As evidenced by studies conducted by MIT’s Sloan School of Management, approximately 85% of the world’s top executives believe that AI will assist their businesses in gaining and sustaining a competitive advantage for years to come. But it’s not just general sentiment; according to studies conducted by PwC, it is anticipated that by the year 2030, AI will have contributed at least $15.7 trillion to the global economy, compared to what it was in 2017 of over $80 trillion. Therefore, it is critical now more than ever before, to begin reaping the benefits from investments in AI.
There are three main fields of AI that are widely used in business applications right now: data analytics, natural language processing, and automation. AI applied to data analytics will effectively mitigate many of the tedious and time consuming processes. Essentially, machine learning will streamline data profiling, matching, and cleaning processes, which will significantly ease the burden held by analysts when preparing data for analysis. Natural language processing is effective in developing intelligent search engines, better accessibility for the visually impaired, and helpful chatbots. These processors are useful in various ways, particularly in assessing customer satisfaction on social media platforms as well as analyzing earnings call transcripts for words that might signify positive or negative outlooks. Finally, automation works to mitigate unnecessarily repetitive processes as well as avoid potentially unsafe tasks. According to research conducted by McKinsey on greater than 2000 work activities across more than 800 different occupations, some jobs are much more easily automatable than others. Many of these include data processing as well as physical work in highly structured and strenuous environments. Although automation will greatly affect the vast majority of jobs in the future, only 5% of occupations right now could be completely automated. Yet, of about 60% of existing occupations right now, McKinsey reported that at least 30% of their activities can be automated. In other words, assuming the rapid implementation of AI in the next two decades, humans will be working alongside machines on a daily basis. We could not reach the level of AI we’re at today had it not been for the massive transformation in data analyzation that has been a major part of business for the past 15 years. Major enterprises have focused efforts on amassing enough data to warrant and utilize AI, and this data has been structured in a way that allows machines to feed experiential formulas to ensure that the integrated AI is based on experiences as opposed to static programming. These developments in big data, in effect, have created the perfect opportunity for AI implementation. The bottom-line is that AI is all about a new type of competitive advantage at an enterprise-scale level. It opens the possibility for companies to develop compound network effects, which stem from AI’s unique ability to automatically and continuously better its performance.
There are a few foreseeable ways in which AI will have a substantial effect on businesses in the near future. First, AI will give large corporations the ability to adopt a completely streamlined supply chain network due to the expansion of inventory-taking drones, program-guided vehicles, and anomaly-detection software. AI will essentially make procurement management less fallible, helping businesses to adapt with the increasing consumer demands in the coming years. Second, AI will help to optimize interactions with customers as well as job applicants by providing additional support processes. These include sentiment analysis technologies as well as live chat software to help businesses gauge interest and general sentiment through psychometric analysis and technologies that evaluate micro-expressions through body language. It is instinctual, of course, for us humans to analyze another’s body language and facial expression, but sometimes we miss things due to our misperceptions and/or our personal biases. AI, after learning from countless past experiences, has the potential ability to read a person’s attitude with greater precision than people in many instances. Cameras can track precise muscle movements, which are then screened by AI systems to analyze its significance, if any. Finally, AI will be instrumental in fighting against more sophisticated cyber breaches. These AI technologies have the unique ability to self-adjust by continuously scrutinizing and learning new data patterns over time. Implementing AI into business will only become increasingly critical to effectively deliver business value and protect confidential data.
At this point, we know many of the amazing opportunities provided by AI to address some of the most specialized customer demands as it can be effective in finding solutions to some of society’s most pressing challenges. However, in a discussion about AI, it’s necessary to consider its disadvantages in order to be ready to offer solutions if problems were to arise because at the end of the day, technology is not infallible. With a greater influence in the lives of humans, AI has an increased responsibility to consider the ethical components of its decisions. In what ways, if at all, can we ensure that AI effectively treats everyone fairly? It’s a question that becomes increasingly pertinent to the implementation of AI in business, yet I’m not totally sure we have the technological means of monitoring its ethical aspects just yet. Due to the fact that AI systems are ultimately designed by humans for human use, they are only as good as the data that they use. Even when one algorithm generates another by an AI system, the original algorithm was potentially subject to some sort of discrete human bias. Therefore, it is critical that we figure out ways to essentially ensure an ethical AI.
Introduced by Swedish philosopher Nick Bostrom, there is an intriguing thought experiment known as the “paperclip maximizer,” which demonstrates the possibility of AI as a serious existential threat. This theoretical example is based around a seemingly innocent activity, creating the greatest number of paper clips possible, just to emphasize a point. He contends that if a paperclip maximizer AI bot had only one goal, to maximize the number of paperclips in its possession, it would work to improve its own intelligence, not for the value of becoming intelligent, but for the sole purpose of figuring out how to collect more paper clips. Doing so would effectively foster its self-improvement, and grow its intelligence by an unprecedented amount incapable by any human. Bostrom surmises that it would transform other machines into paperclip-manufacturing systems, and eventually start destroying civilization itself all in the pursuit of simply doing what it was told to do: make paper clips. The whole point of this thought experiment is to show that in order to truly ensure an ethical and purposeful AI, it’s critical that it has a complete understanding of human values, such as life, love, happiness, variety, etc. so that it effectively understand what both its goals as well as limits are. If not, an overwhelmingly problematic disconnect between the values of us humans and the values of increasingly advanced and intelligent AI systems can quickly develop.
At the end of the day, we need to ensure a balanced level of cooperation between machines and people in regards to the implementation of AI in the future. In my opinion, we cannot simply let AI systems completely take over the work that we do. I believe we must develop technology that can supplement our work, predict various outcomes, consider uncertainties at a large scale, and let humans make the final decision. This is in stark contrast to another field of AI, known as deterministic AI, which encourages AI to reach final decisions completely independent from human intervention.
It’s cliché to say that business moves fast. It moved fast 10 years ago, and it’s moving even faster now. Today is the day of big data at our fingertips, of people speaking to thousands of others in seconds, of instant bank transfers, of next day delivery. Businesses have had to adapt to keep up with the moving times, and will have to continue to do so.
The possibilities are endless when it comes to the development of AI, and there’s a bright future ahead of us in this field. Yet, we must use it responsibly. Tech is not infallible, but it really doesn’t need to be. It has been created to augment and improve, not completely replace. We are in the midst of a fast-paced data renaissance, and I’m thrilled to be a part of it.
Continue the conversation with Ryan at rjf65@georgetown.edu