01Limitations of A.I. Systems
All A.I. systems, as of 2024, are narrowly trained — they may excel in specific tasks but fail in others. Large Language Models can produce “hallucinations,” generating inaccurate or misleading information. Service providers must ensure users understand these limitations and avoid fostering unwarranted trust.
02Transparency About A.I. Processes
A.I. systems are trained on vast datasets that may contain inherent biases. Organizations must be transparent about the processes underlying these systems, clearly label AI-generated content, and inform users of potential risks and limitations.
03Accountability
Human operators must be held accountable for the behaviour and outcomes of A.I. systems. Organizations may delegate tasks to A.I. tools, but cannot delegate responsibility for the ethical and proper execution of those tasks.
04Capacity Building
All citizens must be equipped with digital, media, and information literacy skills to engage with A.I. Fostering critical thinking, creativity, and a questioning mindset helps individuals avoid automation bias and develop informed trust.
05Data Privacy
A.I. systems must adhere to the data privacy principles of the Jamaica Data Protection Act (2020), which emphasizes transparency, consent, and access rights in the handling of personal data.
06Inclusivity and Accessibility
A.I. technologies should be inclusive and accessible to all Jamaicans. Vulnerable and marginalized groups must not be left behind — advancements should bridge digital divides and promote social inclusion.