AI: Bane or Boon

Artificial intelligence (AI) is the talk of the world. It is the new big thing and the next bigger thing as it rapidly evolves from Artificial Narrow Intelligence (ANI) to, probably, Artificial Superintelligence(ASI). It is rapidly transforming almost everything that we have grown so used to. If the development trajectory continues,  AI will transform the way we think, communicate, relate, work, and live.

AI is also changing the way we pursue security and peace, conduct warfare, and manage national power. It is redefining the parameters of global stability and helping identify where conflict escalation can emerge. It drives autonomous weapons and helps drones and planes navigate air defense systems. Thus, we see the dash, mad or rational, to AI dominance.

 It is transforming global and national economies. It is reshaping how wealth is created and distributed, determining which industries, companies, and people become rich. It is changing the competitive positions of industries and companies while altering the skills, capital, and infrastructure that determine economic success.

To financial and tax authorities, for example, AI is like a “liquid” that cannot be held in one location. They describe AI as hyper-mobile, borderless, and impossible to contain within traditional geographic borders.

By every meaningful measure, AI is ushering a bigger tsunami of change than any previous human invention - electricity (probably), printing press, automation and robotics, aviation, nuclear power, and medical breakthroughs. This paradigm-shifting human invention is not only big in its highly anticipated scale of impact, but it is also spreading like wildfire even before it and its implications are fully understood. While previous transformative technologies took years or decades to diffuse through society, AI is compressing the adoption curve dramatically.

It took at least four decades for electricity to be adopted in homes across America and three decades for computers to become personal and household tools. In contrast, ChatGPT, an AI, for example, reached 100 million users in two months. It is estimated that 41 to 43 percent of employees in public and private sectors (Gallup survey) are now using AI tools at work. Yet only  26 percent of government organizations have officially integrated AI in their organizations, creating a “Shadow AI” challenge where employees take it upon themselves to use AI tools to do their jobs faster, bypassing IT and security policies.

The Rushing AI Tsunami of Change

What should we make of the extraordinary wave of change that artificial intelligence is bringing into our workplaces, lives, and yes, to the world and beyond? We are already feeling its first powerful waves through technologies such as ChatGPT, Claude, Gemini, Copilot, Alexa, and Rufus. And this is just the beginning.

So, how should we respond?

Should we be afraid and run as fast and far as we can? Should we duck for cover then swim when the turbulence has passed? Or should we move toward the wave – understand it, embrace its possibilities, and learn to ride it?

Quick Scan of the Benefits of AI

Whether we know it or not, we are already experiencing the convenience, speed, and benefits of AIs. AIs are already at work as conversational assistants and chatbots in virtual assistance, smart home control, and customer service. They are at work in e-commerce and personalization, e.g., product recommendations closely aligned to our preferences following us  as we browse our computers. They may have written and edited the email or meeting invite that you received or the minutes of the meeting you are reading. They are at work in fraud detection and credit scoring. They are at work in routing traffic and may even be driving the car that just passed by. They are at work discovering cures for cancer and other diseases at stunningly faster speed than ever before and speeding up scientific discoveries.

And perhaps most significantly, AI has the potential to dramatically amplify human productivity and our ability to solve complex problems.

Dramatic increase in productivity has been observed at the task level but not as much at the company or economy-wide level. The disparity in the  AI productivity impact at the micro and macro level is what economists call the AI Productivity Paradox.

Scanning the Risks Associated with AI

Yet, as with every transformative technology, AI brings not only extraordinary opportunities but also serious risks.

Perhaps the most immediate and personal concern is economic and workforce disruption. People are scared they may lose their jobs to AI, or they will not be able to find jobs. In the US, it was reported that 175,796 job cuts took place since AI displacement began in late 2022. Because of people’s instinct to see and exaggerate the negative (the so-called negative bias), many focus on job losses due to AI.

However, AI adoption creates jobs too! The World Economic Forum (WEF) projects that AI adoption would create 170 million new roles globally by 2030, resulting in net gain of 78 million jobs. According to the Economist, one million jobs have been created due to adoption of AI in the US, outpacing roughly 200,000 layoffs. AI adoption would create new jobs such as machine learning researchers, data annotators and labelers, prompt engineers, AI enablement leads and architects, AI ethics and compliance officers, data center operators, vibe coders, and non-technical programmers.

In every technology-driven inflection point in history, people lost their jobs. People lost their jobs during the industrial revolution,  and during the introduction of automation, cars, and computers. But after the initial job losses in the transition, employment became stable and even higher. Technology has consistently demonstrated in history it is a net job creator.

Although widely described as “ cold-blooded, calculating machine,“ AI systems can be biased, discriminatory, and erroneous in its decisions, e.g., in who are eligible to a service. These can be due to the algorithm given it, which is like the preconceptions that biased and discriminatory people picked up when they were young. These could also be due to the data it was trained on or fed, which is like the bias and injustice that well-intentioned and fair-minded people commit because of fake news or false information. Garbage in garbage out! And, yes, AI hallucinates too!

AI can generate, facilitate, and amplify misinformation and social manipulation (deepfakes and social bots), cybersecurity breach, biological hazard, bioweapon proliferation, data privacy breach and surveillance, and critical infrastructure dependency. At the wrong hands, they pose serious threats.

Some of the risks can pose existential dangers to humanity. Especially stoke by movies like The Matrix, Terminator, 2001: A Space Odyssey, and I, Robot, some of the risks can scare people.

Even if people are told that AI development is at its early stage of Artificial Narrow Intelligence and at least three decades away from Artificial  Superintelligence, if ever we get there, some people are scared,  with others even citing religious beliefs as grounds for their fear. It does not help them even if they are told by superforecasters in a landmark forecasting study that a nuclear event is three times more likely to end humanity than AI. The probability of human extinction due to nuclear events is put at about 1.1 percent and 0.38 percent, the so-called p(doom), due to AI.

Black, White, and In-between at the Crossroad

The decision that decision-makers have to make is a not clear-cut binary choice of whether to adopt AI or not. The benefits of AI are not unfounded promises, and they are too great to ignore. But neither are the risks. Some are too consequential to ignore. Yet, whether adopting AI or not at global, humanitarian level cannot be fully decided by a cost-benefit analysis because of qualitative, existential, and ethical considerations.

Yes, adopting AI can be made but. And what are the buts? These are the risk mitigations that can be taken like policy and technical guardrails, such as  enactment of  AI policies and regulations, requiring the use of explainable AI(XAI), bias auditing and data curation, mandatory watermarking and provenance tracking, human-in-the-loop mandates, adoption of kill switch as a last resort, and career re-skilling and augmentation training. Personal mitigation measures can be adopted, like turning off training settings, not sharing classified and sensitive information, and turning off options allowing personal data to be used in model training.

AI is called “artificial” because human created it. People are the creators and AIs are the created. Metaphorically, as the creators, people “breathe” or “implant” in AIs algorithms which animate and guide AIs even in their learnings and adaptations. An algorithm defines and frames the development that AI can do on their own. Whether an AI becomes a Frankenstein or not, it is not entirely in its power to make. It is constrained by human perfection, limitations, values, and desires. It depends on how it is designed, instructed, trained, deployed, and monitored. In other words, it is human brokenness, depravity, and ineptitude that create monsters in AIs.

AI, thus, can be a bane or boon. It can make the world a “heap of ruins” or a thriving land of abundance free from dreaded diseases. To George Orwell’s fans and fans of apocalyptic movies, AI can be exciting or scary, but to people with terminal illnesses like cancer, AI can be their only last best hope. What about you? Is AI bane or boon?