Over the past 12 months, public discussion of artificial intelligence in Fiji has remained limited. The main focus has been on the risks that AI poses to social life and workplace productivity. A more structured and balanced discussion that considers both the benefits and risks of this technology is yet to take place.
As you sit in your living room in Fiji in the evening, AI is all around you, unnoticed. A mother on Facebook, whose name she won’t reveal when asked, had her feed arranged by an algorithm.
My daughter uses ChatGPT to complete her 12th grade economics essay in less time than it takes to boil a pitcher of water. The lady next door wants to write on a free app. facebook I’m posting this for her market stall this weekend. My cousin is half watching. YouTube Videos selected by another algorithm. If you ask them if they use artificial intelligence, everyone will say no. Subconsciously, they all exist in different forms.
This is the current state of AI in Fiji. This is not the movie version published in the international press. More quietly, foreign software has flowed into our phones, classrooms, small businesses, and the family group chats that bind us together.
The question we have yet to seriously address is whether this technology will shape how it functions in our lives, or whether we will continue to use it without asking who builds it, who benefits from it, and who pays the bill if it causes harm.
AI has quietly arrived in our daily lives
Most of us in Fiji already use AI every day. There were very few people I could name when we started. When scrolling through TikTok, dictating text messages, automatically translating Messenger chats, or watching videos, facebook Reels are powered by recommendation software, and all of their actions are mediated by machine learning. These are no longer optional features. This is the floor on which the platform stands.
The more visible tools (ChatGPT, Gemini, Copilot, Claude) are built by a few very large foreign companies. OpenAI, Google, Microsoft, Meta, and Anthropic. None of them are based in the Pacific. There is nothing iTaukei, Fiji, Hindi or designed with our economy in mind. That’s not a complaint. This is true and should shape the way we discuss AI in Fiji. Because you can see where design choices are made and where they aren’t.
Misuse we are already experiencing
The damage is no longer hypothetical. Deepfake images and cloned voices are being used for romance scams, investment fraud and political smearing across the Pacific. AI-generated photos build fake Facebook profiles targeting elderly relatives and children. Utilizing the same generative technology, TikTok and Snapchat’s beauty filters continue to cause tangible damage to teens’ self-esteem. Research on this topic is now too extensive to ignore.
Cyberbullying in our middle school was already a serious problem before this incident occurred. Now it’s even worse. AI tools can fabricate persuasive messages and doctored images in seconds. Additionally, existing cybercrime and online safety laws were drafted before this technology existed. They are not suitable at this time. Lacking up-to-date laws, better trained police investigators, and clear public reporting channels.
Students, ChatGPT, and deeper questions
Let me be candid about what I see in my own classroom. Chat GPT It is currently in routine use at USP, FNU, the University of Fiji and secondary schools. Some students use it to write entire essays. Others are using this tool in the way I expected. That is, as a lecturer explaining a concept in three different ways in the middle of the night when the lecturer is not present. Some teachers have secretly started using it to grade assignments. That’s an argument in itself.
Banning the tool doesn’t work. We tried it on our phones, social media, and Wikipedia, and each time technology won.. So the question is not whether students should be allowed to use AI. The question is whether we are teaching them to think critically. When a student in my class asks Chat GPT Whose version of history does she give to explain the events in Suva in May 1987? If you ask her for advice about healing, will the model recognize that knowledge, or will she confidently invent something that seems right? If she asks her to write a poem in Fijian Hindi, will she write a poem, or will she write a clumsy approximation that would be unacceptable to a Fijian Hindi speaker?
These are not abstract concerns. I showed them to my students every semester. The answer is a solemn one. The person who passes the course is not the person who wrote the smartest prompt. They are the ones who learn when to trust the machine and when to push it back. That’s a different skill. We are just beginning to teach it.
Hidden costs behind the screen
Behind the ease of using AI tools on your phone are two costs that rarely surface in marketing. The first is humans. to make Chat GPT and similar products are safe enough to release, yet the companies that make them employ workers in Kenya, the Philippines, and other low-wage economies to read and label the most offensive content on the internet. These workers are paid a few dollars an hour so that wealthy clients don’t have to absorb the psychological damage.
The second cost is environmental. Training a single large language model consumes enormous amounts of power and fresh water for cooling, on a scale that would be surprising to anyone living in the Pacific Ocean. As the use of AI expands, so will our carbon footprint. This is not an incidental problem for a country already plagued by rising sea levels and powerful cyclones. Neither problem can be solved by ourselves, but our silence in local forums is considered consent.
classroom example
Imagine what could happen in any 12th grade classroom this week. Students are asked to write an essay about its cultural significance. Tabuawhale teeth. She types the question as Chat GPT And within seconds, you’ll have three pages of fluent prose. The writing is beautiful. It can also be quietly wrong in some places. Models can cause confusion Tabua Similar items from Tonga are used to explain the ritual in a way unrecognizable to Fijian elders. It almost certainly misses the spiritual weight of the thing because that weight is carried in the slow interactions of our community and village life. The model was trained not on the open internet.
If she submits her essay without asking questions, she will have learned almost nothing. Worse, she had been quietly taught that foreign machines knew her culture better than her grandparents. But the same exercise becomes more powerful when teachers invite the class to compare the AI text to what the elder actually says. Students understand in real time what tools can do, what they can do wrong, and why it matters who conveys cultural knowledge in the first place. In my opinion, that is what AI literacy in Fiji should look like. I don’t remember the warning. Judgment practice.
what actually works
Writing only about harm is dishonest. AI is working for good in Fiji too, and we should say so. Farmers can take pictures of diseased leaves and receive a diagnosis through a phone app. This may be incomplete, but it’s faster than waiting for an extension agent. Nurses on remote islands can use AI-assisted screening to detect diabetic retinopathy, which is becoming more prevalent among the population. A small tour operator in Nadi can operate a multilingual chatbot overnight to respond to inquiries in English, Mandarin, and Japanese without paying for 24-hour staff. A young woman in Labasa can teach herself coding, graphic design and accounting from YouTube and start freelance work from home.
These are not futures. They are happening now. The Fijians who will do well in the next decade will not be those who fear and avoid AI, nor those who embrace it without question. They will be the ones who use it with skill and judgment. The challenge for the rest of us, including those in universities, is to ensure that more young Fijians, not just in urban centers but in the maritime sector and inland, have access to the internet, devices and the training to join them.
Who is AI created for?
Now comes the difficult part. AI is not built equally for everyone. The large-scale language models that most of us use are trained on English data that comes overwhelmingly from Western sources. They understand a New York lawyer better than a teacher in Lautoka, and that’s by design, not by chance. The training corpus contains far more of the former.
Our Pacific languages (iTaukei, Fijian Hindi, Rotuman, and the languages of neighboring countries) are spoken by relatively few people worldwide. As such, they are of no commercial interest to companies deciding what to build next. If we do nothing, the next generation of Pacific children will grow up using technology that does not speak their first language and does not know their stories. It’s not a small cultural concern. Over time, who we are slowly erodes.
The answer is not to deny AI. The answer is for Pacific institutions, myself included, to invest in Pacific-led research. Datasets are written in our own language, built with community consent, and used to train tools that serve our people. Iceland has achieved this with the Icelandic language. Aotearoa’s Māori communities are doing this Teleo. There’s no honest reason why it can’t be done.
what must be done
Concrete actions are due at four levels. Within the next 12 months, the Fijian government must publish a national AI strategy, drafted after serious consultation with educators, businesses, faith communities and young people, and covering minimum data protection, cybercrime, education and language preservation.
The Department of Education should work with USP and other universities to issue clear classroom guidelines that focus on critical literacy skills, rather than a blanket ban.
The private sector, especially tourism, agriculture and creative industries, should invest in AI training for their workforces, with the support of development partners.
And the rest of us as parents, employers, neighbors, and citizens need to understand these tools, too. It’s not about mastering them. Just be knowledgeable enough to ask smart questions when your kids or colleagues use them.
Choice, not waves
It’s useful to describe AI as happening to us like a tide. This convenience is what matters, and Tide is not responsible for anything. But AI is not a trend. It is a long series of human decisions, made every day by people in every office you can name, investigating and contesting interests. Fiji has a small population.
Our voices in Pacific forums, international organizations and at home are authentic.
We can choose to be informed and organized. We can also choose to be quiet. There is no third option. If we want our children to grow up as thinkers and builders rather than customers of foreign software, we must start doing so now.
Dr. Sharma is a senior lecturer at the USP School of Management. The views expressed here are his own and do not represent the views of his employer.
