Hong Kong payments firm YeahPay explores agentic payments

Hong Kong’s real concern regarding artificial intelligence (AI) is not as dramatic as many portray it. While AI has captured public imagination as an escape from labor-intensive tasks and a potential catalyst for a new industrial revolution, the reality is less sensational. Governments are racing to draft AI strategies, businesses are adapting their operations, and investors are pouring resources into AI infrastructure.

However, there is considerable uncertainty about the long-term trajectory. The Monetary Authority of Hong Kong recently examined a major bank's decision to locate its AI center in Singapore and hire over a hundred AI specialists. This illustrates the competitive pressure at play, but establishing AI centers does not automatically address the accompanying risks.

Investors might accept some uncertainty as long as financial returns are realized, but regulators cannot. Understanding the potential impact of AI is a distinct decision from merely incorporating it into operations. Simultaneously, leading AI developers have expressed concerns about the technology's potential dangers. Ironically, these very developers are contributing to the advancement of AI.

While there is a conflict of interest, it does not negate the validity of their warnings. However, simply hearing a warning is not enough. Companies must adopt AI tools cautiously to avoid importing external perspectives that may be flawed. Hong Kong’s government is addressing the legal implications of AI use, with the Department of Justice leading an interdepartmental review of relevant legislation.

This is a step in the right direction, but equipping individuals to grasp the nuances of AI is equally crucial. Even legal professionals require supervision to verify AI-generated outputs. For the average person, the task of interpreting AI-generated results is daunting. AI can augment existing threats, but its transformative impact on decision-making processes necessitates new safeguards.

For instance, if an AI agent unknowingly signs a contract on a principal’s behalf, the contractual binding remains a fresh issue. The notion of an AI apocalypse is undeniably dramatic, but the actual challenge is more subdued: preventing the gradual delegation of judgment to systems that may not always be reliable.