For the past years I have been doing a lot of reading on AI. Not technical reading but reading related to its likely economic and social implications. The more I read, the more I recall a part of the lyrics from Sting's 1993 hit "If I Ever Lose My Faith In You" which said "I never saw no miracle of science that didn't go from a blessing to a curse." These resonant words feel eerily prophetic when viewed through the lens of artificial intelligence. Today, we stand at a similar juncture, witnessing a technological marvel hailed by many as an unmitigated force for good, yet whispered about by others as a potential harbinger of unforeseen challenges.

The dominant narrative, amplified by leading international and local articles, often focuses on AI's transformative potential, particularly concerning increased productivity. Financial markets, ever-eager for the next growth engine, have largely embraced this optimistic outlook. Investment in AI has surged, with global spending projected to exceed $500 billion by 2027, and countless analysis point to AI's potential to add trillions to the global economy. Companies integrating AI are seeing significant boosts in efficiency, streamlined operations, and new avenues for innovation, fuelling a widespread conviction that financial returns are almost a given. The NASDAQ and other tech-heavy indices have experienced notable rallies driven by AI enthusiasm, reflecting investors' unwavering belief in its positive economic impact.

The benefits of artificial intelligence are indeed vast and well-documented across numerous fields. Leading research articles and books consistently highlight its capacity to enhance human capabilities and solve some of the world's most intractable problems. In healthcare, AI is revolutionizing diagnostics, drug discovery, and personalized medicine. Algorithms can analyse medical images with greater accuracy than human experts, identifying diseases like cancer and retinopathy in their earliest stages. Research published in Nature Medicine frequently showcases AI's role in accelerating drug development by predicting molecular interactions and optimizing clinical trials, drastically reducing time and cost. AI-powered tools assist surgeons, predict patient outcomes, and even provide mental health support through intelligent chatbots, making healthcare more accessible and effective.

Scientific research is experiencing an unprecedented acceleration. From climate modelling to materials science, AI can process and interpret colossal datasets that would overwhelm human researchers. AlphaFold, a programme developed by DeepMind, has made groundbreaking progress in protein folding prediction, a challenge that has eluded scientists for decades, potentially unlocking new treatments for diseases. This kind of AI-driven discovery is not just incremental but foundational, changing the very methodologies of scientific inquiry.

Beyond these high-impact areas, AI is driving efficiencies and fostering innovation across virtually every industry. In manufacturing, predictive maintenance systems powered by AI prevent costly breakdowns and optimize production lines. In logistics, AI algorithms optimise delivery routes and inventory management, reducing waste and improving efficiency. Even in creative fields, AI is assisting artists, musicians, and writers, not replacing them, but providing new tools and perspectives. Leading academic journals like Artificial Intelligence and IEEE Transactions on Pattern Analysis and Machine Intelligence consistently feature studies detailing these diverse applications, underscoring AI's profound positive impact on productivity, problem-solving, and human welfare.

Despite the dazzling promises, a growing chorus of leading researchers, economists, and philosophers warn of significant dangers lurking within the rapid advancement of AI. Their concerns move beyond mere technological glitches to fundamental questions about human agency, societal structures, and the very nature of intelligence.

One of the most compelling critiques comes from Daron Acemoglu (Nobel laureate), co-author of Why Nations Fail and Power and Progress: Our Thousand-Year Struggle Over Technology and Prosperity. Acemoglu argues that without proper societal direction, AI could lead to what he terms "so-so technologies," which primarily serve to automate tasks, deskill workers, and consolidate power rather than truly augment human capabilities or create new, high-value jobs. He highlights the historical pattern where technological revolutions, while increasing overall productivity, often lead to a greater share of the gains accumulating at the top, exacerbating inequality and eroding the bargaining power of labour. His work underscores the critical importance of ensuring that AI development is guided by democratic principles and a commitment to shared prosperity, rather than solely by market forces.

Similarly, Yuval Noah Harari, author of Sapiens, Homo Deus, and 21 Lessons for the 21st Century, offers a profound and often unsettling perspective on the existential threats posed by advanced AI. Harari's primary concern revolves around the potential for AI to undermine human free will, reshape our societies, and even create a new, potentially non-human, dominant intelligence. He vividly describes a future where algorithms know us better than we know ourselves, manipulating our choices and preferences through hyper-personalised content and nudges. More chillingly, Harari warns of the "data religion," where human experience is reduced to data points, and the authority of algorithms supplants traditional sources of wisdom and moral guidance. His work emphasises the profound ethical and philosophical challenges that accompany AI, urging humanity to consider the very definition of consciousness and purpose in an increasingly algorithmic world.

A particularly insidious peril highlighted by these and other scholars is deskilling, especially concerning our thinking skills. As AI tools become more sophisticated, taking over complex analytical and problem-solving tasks, there's a legitimate fear that human cognitive abilities could atrophy. If critical thinking, complex reasoning, and creative problem-solving are increasingly outsourced to machines, will humans become less capable of these essential intellectual functions? For instance, while AI can write eloquent summaries or even generate ideas, the constant reliance on such tools might diminish our capacity for original thought, deep analysis, and the arduous process of wrestling with complex problems independently. This intellectual atrophy could have profound implications for education, innovation, and our ability to navigate future crises without external algorithmic guidance. This is not merely about job displacement, but about the very erosion of what makes us intellectually human.

The dangers also extend to biases embedded in algorithms, which can perpetuate and even amplify societal inequalities; the weaponisation of AI in autonomous systems; the erosion of privacy through ubiquitous surveillance; and the potential for "black box" AI systems whose decisions are opaque and uncontrollable, leading to a loss of accountability.

In the face of these dichotomous perspectives, the European Union has taken a bold and unprecedented step with the AI Act. This landmark legislation, the first comprehensive legal framework for AI globally, aims to regulate AI systems based on their potential risk levels. Its core intention is clear: to foster trustworthy AI that respects fundamental rights, safety, and ethical principles.

The Act categorises AI systems into different risk levels - from "unacceptable risk" (e.g., social scoring by governments) which are banned, to "high-risk" (e.g., AI in critical infrastructure, employment, law enforcement), which face stringent requirements for data quality, human oversight, transparency, and robustness. "Limited risk" AI systems have lighter transparency obligations, while "minimal risk" AI (like spam filters) are largely unregulated.

While the EU has faced criticism for potentially stifling innovation, with some arguing that its stringent regulations could push AI development and investment to other regions like the US and China, its approach reflects a strategic long-term vision. The EU understands that while rapid innovation is desirable, unchecked technological advancement can lead to significant societal costs. By setting a global standard for ethical and trustworthy AI, the EU is playing the long game. It aims to prevent the very perils highlighted by Acemoglu, Harari, and others - the exacerbation of inequality, the erosion of human agency, and the deployment of potentially harmful technologies.

The EU's bet is that in a future where AI is ubiquitous, the market will increasingly value AI systems that are not only powerful but also transparent, fair, and accountable. This regulatory framework is an attempt to assure that the "blessing" of AI does not swiftly transform into a "curse" for its citizens and to position Europe as a leader in ethical AI deployment, thereby shaping the global norms for this transformative technology.

Artificial intelligence stands as a testament to human ingenuity, a true miracle of science with the power to unlock unprecedented progress. Yet, like all profound scientific advancements, its trajectory is not predetermined. It can indeed be a blessing, pushing the boundaries of discovery, eradicating disease, and fostering unimaginable efficiencies. But without conscious, collective guidance, without the foresight to anticipate its shadows, it risks becoming a curse - deepening inequalities, undermining human dignity, and even diminishing our very capacity for independent thought.

The choice, ultimately, is ours. We cannot afford to be passive observers, merely hoping that the positive aspects will prevail. We must actively shape AI's future. This demands more than just technological prowess; it requires profound ethical deliberation, sensible global regulatory frameworks and a renewed commitment to human-centric design. It means fostering an education system that prioritises critical thinking, creativity, and emotional intelligence - skills that cannot be easily replicated or deskilled by machines.

The responsibility for ensuring that this miracle of science truly benefits all of humanity, without succumbing to its potential for harm, rests firmly on our shoulders. The future is not just about building smarter machines; it's about building a wiser society that wields them responsibly. Let us rise to this challenge, daring to envision and create an AI future that amplifies our blessings and consciously averts the curse.