What do we mean by AI Safety?

Although there is a growing conversation in Hong Kong around navigating increasingly advanced AI, there are major neglected safety risks that are much larger in scale and more consequential.

The MIT AI Risk Repository comprehensively lists over 1700 AI risks, organised into 7 domains.

Amongst these, we note that many major risks are rarely discussed on official or mainstream platforms in Hong Kong, such as:

Disempowerment of humanity, in the sense of society and culture as well as AI takeover of systems like decision-making processes and economic functions,
Impact on users’ mental well-being and cognitive development, especially for children.
Environmental harm, including water and energy usage
Risks to critical infrastructure - potentially catastrophic events for which we’re under-prepared,
Possibly even existential risks to humanity caused by misaligned powerful AI.

We therefore see a critical need to expand Hong Kong’s public discourse on safe development and adoption of advanced AI.

AI Safety Hong Kong is an independent, neutral initiative.
We are not affiliated with any government body or commercial AI provider.

Here's what we're building and how: