Large language models (LLMs) are revolutionizing the world of artificial intelligence, with applications ranging from machine translation to content creation. However, speakers of low-resource languages, such as Caribbean Creole languages, face a significant dilemma in this new era. 
LLMs require vast amounts of data to train effectively. For low-resource languages, this data is often scarce, limiting the models' capabilities and hindering access to advanced AI tools. To overcome this challenge, speakers can choose to make their language data freely available on the web to train LLMs owned by multinational AI corporations. This approach, however, raises ethical concerns. 
 
While these corporations profit from the use of these languages, they often contribute little to the welfare and development of the communities that speak them. Alternatively, speakers can choose to protect their language data, restricting its use and potentially hindering their access to the very productivity tools needed for economic and social development. 
 
This talk will explore this complex dilemma, examining the ethical, economic, and social implications of making language data available to multinational corporations for profit making and the continued exploitation of speakers of Caribbean languages, versus protecting such data from incorporation into the LLMs of the big corporations, to the potential detriment of local communities.