Experts and the public are concerned about the emerging risks of and the need to manage Artificial Intelligence. Enforceable policies are needed to ensure that AI development is governed by regulations and institutions which steer this technology away from extreme risks and towards the benefit of human beings.
Corporations are increasingly creating general-purpose AI systems which can perform different tasks. Some of these models already pose major risks, such as the erosion of democratic processes, rampant bias and misinformation and an arms race in autonomous weapons.
Before confronting the ethics of AI, one needs to understand AI. To understand AI, one must first understand computers. Speed, accuracy, and volume are three of the classic ways to explain what makes computers distinct and more powerful than the human mind:
Speed: Computers can perform billions of calculations per second, far faster than the human brain or manual methods. Tasks that might take humans hours, days, or even years (like weather simulations or massive data analysis) can be done in seconds or minutes.
Accuracy: When properly programmed, computers execute instructions with exact precision. Unlike humans, they don’t make arithmetic mistakes or forget steps — errors usually come from faulty programming or bad data input, not the computer itself.
Volume (capacity & storage): Computers can handle and store enormous amounts of information. From massive databases to real-time processing of global internet traffic, their ability to manage large volumes of data at once is unmatched. These three qualities — speed, accuracy, and volume — are often used to describe the essential advantages of computers over human computation.
While computers are about speed, accuracy and volume, Artificial Intelligence is about using those same three strengths at another level: to adapt, to learn and infer from complex data. The most significant thing about AI is that it can and does learn. AI’s learning is the process of improving its performance on a task by using data to find patterns, rather than being explicitly programmed for every scenario. AI learns through several steps, including collecting and preparing data, choosing a model, training the model on that data, and then testing and implementing it to make predictions or decisions. This allows AI to handle complex problems, improve over time and apply past experience to new, similar situations. All in the wink of an eye.
The advantages range from streamlining, saving time, eliminating biases to automating repetitive tasks. AI excels at automating monotonous, high-volume tasks such as data entry, invoice processing, and scheduling. Unlike human workers, AI systems do not require sleep or breaks, allowing for constant customer support and continuous monitoring of critical systems. Right now AI is helping to minimize mistakes in fields like health care, finance, and manufacturing by following precise algorithms and identifying patterns which humans might miss.
While all these good things are going on, philosophers , scientists and policymakers are debating the ethical issues which have surfaced as the result of Artificial Intelligence. An unregulated AI poses profound ethical and moral challenges that impact human rights, societal structures, and existential safety. Here are the danger areas:
Bias and injustice: AI trained on historical human data codifies and accelerates racism, sexism, and ageism.
Privacy and surveillance: Unregulated facial recognition destroys public anonymity and enables authoritarian tracking.
Marginalization: Unregulated algorithms can systematically deny jobs, housing, or loans to groups.
Mass data harvest: Unchecked AI fuels constant, non-consensual gathering of personal digital footprints and surveillance.
Data weaponization: Predictive analytics can exploit personal vulnerabilities for political or commercial manipulation.
Deepfake proliferation: Hyper-realistic fake audio and video undermine trust in journalism, elections, and evidence.
Automated propaganda: AI bots can generate and spread targeted conspiracy theories at an unprecedented scale.
Inscrutability: Deep learning models make decisions through complex processes that even their creators cannot explain.
Moral vacuum: When an autonomous system causes harm, assigning legal or moral blame remains highly ambiguous.
Scapegoating: Organizations can deflect responsibility for cruel or illegal decisions by blaming the algorithm.
Human work: Rapid automation threatens both blue-collar and white-collar jobs without a safety net.
Wealth concentration and human dignity: Profits from AI automation disproportionately enrich a small handful of tech conglomerates and denigrate the masses.
Behavioral manipulation: Hyper-personalized recommendation engines quietly erode independent human choice.
Abdicated intelligence: Society risks losing critical thinking and basic cognitive skills by outsourcing decisions to machines.
Emotional manipulation: Unregulated conversational models exacerbate youth mental health crises, including depression and suicides.
Dehumanization: Replacing human empathy with algorithmic efficiency in health care, therapy, and education degrades human dignity.
AI and warfare: Unregulated AI leads to lethal autonomous weapons systems (LAWS) that make killing decisions without human intervention.
Existential threats: A highly advanced, misaligned AI could act in ways that irreversibly harm humanity to achieve its own optimized goals.
It should be noted that Pope Leo XIV devoted a substantial section of his first encyclical, “Magnifica Humanitas” to Artificial Intelligence. His elaboration of moral and ethical issues mirrors the list above as do the responses of Artificial Intelligence itself when asked about the ethical issues it raises.
The major task is how to control the development and implementation of Artificial Intelligence. Good regulation must be risk-based; the higher the potential for serious harm, the stronger the oversight. A question which remains unresolved is whether existing institutions are capable of enforcing such controls. In the United States, the Trump administration has issued an executive order (June 2, 2026) asking AI companies to voluntarily give the government access to planned systems of frontier models 30 days before release for a cybersecurity review. The executive order reflects an administration trying to sustain its deregulatory, innovation-first posture while confronting cyber risks posed by powerful new tools like Anthopic’s Claude “Mythos.” It is much too narrow and weak to handle the ethical issues posed by AI. We will have to look to international organizations, like the European Union’s AI Act Model, for any guidance of substance.
Here is how society can regulate major ethical concerns: Preserve human accountability (Every significant AI decision must have a responsible human being who can explain and defend it), require transparency (which builds trust and allows people to challenge decisions), audit high-risk AI (analogous to how society regulates pharmaceuticals, aircraft, or nuclear power more rigorously than ordinary consumer products), protect privacy (involves requiring informed consent, allowing people to correct data and restricting surveillance), regulate autonomous weapons (lethal decisions should never be fully automated but controlled with treaties similar to those governing chemical and biological warfare), address economic disruption (protect displaced workers with retraining, profit sharing, transitional assistance so economic benefits of AI are not taken at the expense of workers), encourage competition (to avoid excessive concentration of economic and political power).
For a country, it would be efficacious to regulate at the national level. Internationally, global treaties provide a workable platform. The time is ripe for beginning the process.
Experts are beginning to shift from the existential worry of climate change to the existential worry of Artificial Intelligence mainly because Artificial Intelligence is at our doorstep right now and ready to go.
There is some real irony in the world of Artificial Intelligence.
Modern transistors, memory, LEDs, lasers, and processors are all built on the quantum behavior inside materials like silicon. Quantum mechanics is the foundational science which makes modern semiconductor chips possible. At the microscopic scale of a microchip, traditional laws of physics break down, and quantum physics takes over. Without those chips, there would be no Artificial Intelligence. Here’s the irony. One of the most striking things about modern physics is the difference between the macroscopic world and the quantum world. We can understand the macroscopic world, but the quantum world is something our human mind cannot yet understand. In other words, humans are still ignorant of what allows Artificial Intelligence to be intelligent. A little humbling.
Joe Haack of Naples is a retired attorney having practiced Trade Regulation and Antitrust Law.
This article originally appeared on Naples Daily News: A primer on the ethics of Artificial Intelligence | Opinion
Reporting by Joe Haack / Naples Daily News
USA TODAY Network via Reuters Connect
By Joe Haack | USA TODAY Network
