Will AI Destroy Humanity?

By Dr Eustache Placide

Artificial intelligence has entered another remarkable phase of its development. Some researchers working at the frontier of AI are now warning that increasingly powerful systems could eventually pose catastrophic risks to humanity. Others argue that such predictions are highly speculative, difficult to quantify and risk distracting attention from the very real problems associated with AI today.

The headlines are dramatic. Could artificial intelligence become so powerful that humans lose control of it? Could increasingly autonomous systems act in ways their creators did not anticipate? Could the technology eventually pose an existential threat? These are legitimate questions for researchers to investigate. But for most of us, there is a more immediate question: how should we govern a powerful technology when its capabilities are developing faster than our institutions, policies and understanding of its consequences? That is a question that should matter to the Cayman Islands.

Recent warnings have made the debate unusually public. Anthropic researcher Evan Hubinger has said that he personally places the probability of AI causing human extinction within the next decade at greater than 10 per cent. Former Anthropic researcher Jacob Coxon resigned while warning that frontier laboratories were moving too quickly toward increasingly powerful systems. These are striking claims, but they are not established scientific forecasts. There is no historical dataset for human-level or superhuman artificial intelligence from which anyone can calculate such an outcome with conventional precision, and researchers disagree sharply about both the likelihood and the timeline.

That uncertainty, however, should not be confused with the absence of risk. We do not need to imagine conscious machines or science-fiction scenarios to identify legitimate concerns surrounding artificial intelligence. AI can already generate convincing false information, impersonate individuals, automate aspects of fraud, assist in cyber activity, and produce text, images, audio, video, and computer code at extraordinary speed. Increasingly capable AI agents can also use tools and take actions rather than answer questions.

At the same time, AI offers tremendous opportunities. It can assist researchers, improve accessibility, increase organisational productivity, support financial analysis, enhance education and allow small businesses to perform tasks that previously required significantly greater resources. That combination of opportunity and risk is precisely why governance matters.

For the Cayman Islands, this conversation has particular significance. We are an international financial centre with a highly connected economy. Our institutions manage sensitive financial, personal, educational and governmental information. Our businesses increasingly depend upon digital systems, while our schools and universities are preparing students for workplaces in which AI will almost certainly become commonplace.

Consider financial services. AI can potentially strengthen fraud detection, compliance monitoring, customer service, document analysis and risk assessment. But its adoption raises important questions. What information is being provided to these systems? Where is that information stored? Can an organisation explain an AI-supported decision? How do we know when the system is wrong? And ultimately, who remains accountable for the decision?

Consider education. Students now have access to AI systems capable of writing essays, generating computer programs, solving mathematical problems and analysing complex information within seconds. Simply banning these technologies would leave students poorly prepared for the world they are entering. But unrestricted use creates the opposite problem: students may gradually outsource the very thinking that education is intended to develop. The challenge is therefore not simply whether students should use AI. It is determining what thinking must remain with the student even when AI is available.

AI governance is sometimes interpreted as an attempt to restrict technology. It should not be. Good governance establishes the conditions under which innovation can occur responsibly. Before deploying AI, an organisation should be able to answer some fundamental questions. What decisions are we allowing AI to influence? What information can employees, customers or students provide to these systems? Where must a human remain responsible? How will AI-generated information be verified? What happens when the system gets something wrong? Who is accountable for the consequences?

The Cayman Islands does not need to wait until every international debate surrounding artificial intelligence has been resolved. Nor do we need to predict exactly what AI will look like five or ten years from now. Instead, Cayman can concentrate on developing the institutional capacity to govern whatever comes next. That means improving AI literacy among leaders, educators and employees; establishing clear policies for responsible use; protecting confidential information; conducting appropriate due diligence before systems are acquired and deployed; and ensuring that humans retain meaningful oversight of consequential decisions.

Higher education has a particularly important responsibility. Universities should not merely teach students how to use AI tools. We should help students understand when to use them, when not to use them, how to question their outputs, how to verify information and how to remain intellectually responsible for work produced with AI assistance. The principle is straightforward: AI should augment human judgment, not quietly replace human accountability.

Here in Cayman, that work is already underway at the University College of the Cayman Islands (UCCI). UCCI has taken a university-wide approach to integrating AI across the curriculum. At the same time, its Workforce Readiness Centre offers AI Foundations: Concepts, Tools & Applications for professionals seeking practical AI skills. The programme explicitly addresses ethical issues, bias and risk, and a recent cohort brought together participants from education, finance, insurance, government, business and data analytics. That combination of practical capability, critical thinking and responsible use is the kind of institutional capacity Cayman will increasingly need as AI becomes embedded across the economy.

Will artificial intelligence eventually become powerful enough to threaten humanity? We do not know. The possibility deserves serious research. But debating the most extreme possible outcome should not distract us from decisions that governments, businesses, educational institutions and individuals are already making about AI today.

For Cayman, a more useful question may therefore be this: how do we capture the extraordinary opportunities presented by artificial intelligence while ensuring that humans remain responsible for the decisions that matter? We do not have to choose between fearing artificial intelligence and embracing it without question. Between fear and complacency lies responsible governance. For the Cayman Islands, that may be one of the most important conversations we can have as we enter the next stage of the AI era.

Author note: Dr. Eustache Placide is a computer science and artificial intelligence professor at the University College of the Cayman Islands. The views and ideas expressed are solely those of the author and do not necessarily represent the positions or policies of UCCI.