Introduction
In recent months, our attention has been focused on new developments in Artificial Intelligence (AI). Particularly with the release of ChatGTP, companies are wondering how they can take advantage of the tool to make their businesses more efficient and innovative. Various authors and academics refer to the opportunities of AI to automate tasks, streamline processes and optimize decision-making, both in the public and private sectors.
Bill Gates, in his note published on March 23, 2023, entitled "The Age of AI has begun," narrated with astonishment how GPT, an AI model, answered 60 multiple-choice questions and 6 open-ended questions from an Advanced Biology (AP Bio) exam, achieving the highest possible score in a university-level biology course. He indicates that at that moment, he understood that he was witnessing the most advanced technological advance of recent times, and that the era of artificial intelligence has begun.
On the other hand, in recent days, a group of technology sector entrepreneurs, experts, and politicians, including Elon Musk, requested a six-month suspension of experiments with the most powerful AI systems, considering that these systems "may pose profound risks to society and humanity."
The purpose of this article is to clarify some concepts. Firstly, what is AI and how does it work? What are the opportunities that AI represents for use in different sectors, particularly in businesses? What are the challenges of using AI? Can we safely leverage AI for decision-making, including structuring businesses?
Definition of AI and Technology
To find a generally accepted definition of AI, I asked GPT-4, the most advanced AI system developed by Open AI to date, the following: What is AI and how does it work? This was its response:
"Artificial intelligence (AI) refers to the development of computer systems that can perform tasks that typically require human intelligence, such as visual perception, speech recognition, decision-making, and language translation. AI works by using algorithms and statistical models to analyze data, identify patterns, and make predictions or decisions based on the information it has learned.
There are various approaches to AI, including rule-based systems, expert systems, neural networks, and machine learning. Rule-based systems use if-then statements to make decisions, while expert systems rely on the knowledge of subject matter experts to solve complex problems. Neural networks are designed to mimic the structure and function of the human brain, and machine learning involves training algorithms to learn from data and improve their accuracy over time.
The field of AI continues to evolve rapidly, with new techniques and applications emerging all the time. As computers become more powerful and data more abundant, AI is likely to play an increasingly important role in many aspects of our lives, from healthcare and education to business and entertainment."
This definition seemed quite comprehensive and accurate, considering the various papers and research works to which I had access in recent studies on AI & Law.
Opportunities
There is no doubt that AI will bring benefits to humanity, considering its enormous capacity to process, organize, classify, and generate patterns and solutions from large amounts of data, in an expeditious manner. This capacity can undoubtedly be used to achieve goals that until recently where impossible to reach using only human processing power.
We have seen how the use of AI and big data has been impacting in different fields. In the financial market, for example, AI has been used to identify investment opportunities, risk management, and decision-making. AI and technology have also disrupted the financial services business, enabling the digital transformation of banks and boosting financial access through fintechs.
In the health sector, AI represents the possibility of having more rigorous diagnoses, automated treatment, monitoring, and prevention, as well as identifying means to improve the quality and extend the human life expectancy. According to the EIT Health and McKinsey report from March 2020 "Transforming healthcare with AI," there is an imminent growth in AI funding for health by venture capital firms and technology giants, as well as clinical trials on the benefits of applying AI. Since 2017, there have already been several FDA approvals for AI in the healthcare sector.
Likewise, the public sector in different countries has been rapidly integrating AI to make public services more efficient. By facilitating data analysis and more accurate predictions, optimal decisions can be made in the management of natural resources, public safety, and urban planning. It has also been analyzed how AI can optimize the way justice is regulated and administered. Moreover, AI promises to secure equal access to education and health for less favored and displaced population in the world.
In the legal field, AI is being used in different tools and applications to predict legal outcomes, expedite and facilitate the drafting and closing of contracts, data processing, and decision making. As in other areas, AI is making easier for professionals to carry out their work, promising a better work-life balance.
Particularly in business structuring and tax planning, the use of AI allows for deeper analysis and more accurate results. As matter of fact, in my thesis published in 2016 on legal strategy for business structuring based on tax planning, I propose a technology-based decision-making methodology with the aim of structuring businesses. This methodology considered the following: (1) assessment of all tax implications of a particular business scenario and determination of possible applicable tax benefits or exemptions; (2) identification of the most efficient combinations of legal, contractual, and corporate tools, according to the particular business objective and scenario; (3) identification of the most appropriate accounting and financial tools according to the business objective; and (4) determination of the most efficient outcome based on a financial projection. The proposed solution is to be achieved by translating analog legal, accounting, and financial data into a digital code or language organized in a way that generates rule-based or even machine learning-based results.
Undoubtedly, AI promises optimal results in business, even better than expected. There are already many existing technology opportunities to leverage for more systematic and precise decision-making. In fact, for the first time, a Chinese company has appointed an AI as its new CEO. Tang Yu, the new CEO of AI, works 24 hours a day, 7 days a week, and receives a salary of $0 (Source: Mundo Ejecutivo, March 21, 2023).
Challenges
AI also brings a series of challenges that are important to consider for its impact on society. One example is the consequences of AI on intellectual property rights, including the value that will be given to human innovation or inventions vs. AI. We need to ask ourselves if AI really has the ability to innovate and be creative, or if it simply uses human invention as a base, and organizes and systematizes it.
In this regard, we have also observed how AI has the possibility of surpassing a human's processing and analysis capacity, and we wonder if consequently, AI has the ability to make more accurate and optimal decisions than a specialized technician or professional with years of experience. There has also been concern that the workforce will be replaced in the not-too-distant future by AI entities, completely disrupting how society operates today. In addition, employees will have to coexist in their workspace and even receive instructions from machines or robots, which will probably entail inconveniences and even occupational risks. Will AI be capable enough to be a constructive leader for an entire organization?
One of the main reasons why I understand that this is not entirely correct is due to the risk of AI bias, particularly in the legal field. AI systems will be as efficient or accurate as the data they are trained on, and if that data is biased or flawed, the system will be too. This can lead to unwanted consequences or unfair results.
Furthermore, those systems created to make decisions based on laws will encounter the problem that the language of the laws is sometimes vague or imprecise or it can lead to multiple interpretations. Likewise, an AI system or model may not be able to put abstract or subjective elements into context or consideration (such as moral or ethical aspects) to ensure a fair or equitable decision or solution, as a judge or lawyer is capable of doing.
Education is already experiencing a major challenge, considering that tools like GPT-4 can perform in-depth research and essays in seconds. Although access to knowledge and information is something to be applauded, it is worrying that these tools will eventually impact, displace or diminish the ability and capacity of human beings to carry out deep analysis and creative tasks.
Likewise, there is discussion about the unpredictability and autonomy of certain AI systems, as well as the inability to explain the results of AI decisions, which brings uncertainty and a lack of confidence that in all cases AI models will act in the best interests of humanity and its environment.
Another interesting discussion arises from the conceptual challenge of assigning moral and legal responsibility for damages caused by autonomous systems or machines that are in the process of learning automatically (machine learning). The legal system will find it challenging to manage the consequences of the increased use of AI and the public risk, especially ensuring that those who have been affected by damages or harm caused by the action of an AI system are fairly compensated.
The lack of regulation and uncertainty as to whether there are sufficient controls by the companies developing these types of models is also a concern. In this regard, history has shown that regulation appears as damages occur or are imminent, thus humanity assuming all kinds of risks and damages in the pre-regulation period.
Conclusion
AI definitely brings many opportunities for humanity, particularly to make processes and decision-making more efficient for both the public and private sectors. However, considering the multiple biases and challenges mentioned, it is prudent for experts, academics, governments, international organizations, and relevant NGOs to join forces to conduct a deep analysis of the opportunities vs. risks of AI and immerse themselves in a responsible discussion of how to face and mitigate these risks in the near future, without discouraging investment and development of AI for the benefit of societies.
It is surely necessary to identify which minimum regulatory elements are reasonably required to ensure sufficient safety controls. Similarly, the liability system in the different jurisdictions should be reviewed or adjusted when necessary, so that liability is fairly assigned and the companies that develop these AI systems are incentivized to apply adequate security measures. For these purposes, world leaders and policy makers must strive to understand AI and how this technology works so efficient solutions are soon established. Time is of essence, considering the speed with which AI development is going.
Finally, I share the opinion of several authors that AI will not displace professionals and specialists such as lawyers and doctors for at least a good while, considering the inaccuracies that can often result from the work of an AI. However, professionals who do not use technological tools and AI to optimize their processes and decision-making will be likely to be displaced by those who do. The above, considering the competitive advantages that technology and AI represents to automate processes and use data to prevent or predict outcomes.
