Literacy Before Law: The Bahamas Bets Its AI Strategy on Training First, Legislation Second

Most Caribbean governments write an AI policy first and worry about whether anyone understands it later. The Bahamas has inverted that order on purpose. Over eleven weeks this summer, Minister of Innovation and National Development Sebas Bastian put a committee, a budget, and a public argument behind a single bet: that citizens should learn to use AI before the government asks them to live under a law written about it. It is an unusual sequence, it is not yet proven, and it now has enough money and enough named institutions behind it to be worth checking against evidence rather than taking on faith.

The evidence exists closer to home than the Bahamas government seems to have looked. A regional training study, a small grassroots pilot two islands away, and a national adoption survey published this month all bear on the same question: does teaching people to use AI first actually change what a country does with it, or does it just make the eventual law easier to sell.

• Sebas Bastian announced a National Artificial Intelligence Committee on June 15, 2026, tasked with drafting the Bahamas' first AI legislation, ahead of a training investment that would arrive two months later.
• On August 13, 2026, the government confirmed $3 million in AI skills training, funded through an Inter-American Development Bank loan and delivered with the provider General Assembly, split between basic literacy and advanced product-building instruction.
• On September 4, 2026, Bastian framed a National AI Literacy Initiative around trust rather than technology, arguing Bahamians need to become creators in the AI economy rather than only consumers of tools built elsewhere.
• Barbados already ran a small version of the same idea. On September 1, 2026, the National Organisation of Women launched free AI training for up to 35 women built around a specific business outcome, not general literacy.

A Budget Speech Committee, Then Silence

Bastian's June 15 announcement, made during the government's 2026/2027 Budget Debate contribution, was specific about structure and vague about timing. The National Artificial Intelligence Committee would bring together public sector officials, private sector representatives, and members of the legal community, backed by a wider advisory forum spanning business, technology, academia, and civil society. Its job is to draft legislation grounded in Bahamian values covering privacy, cyber security, transparency, and responsible use, with public consultation built into the process rather than bolted on afterward. What the government has not published, as of this article, is a reporting deadline, a consultation schedule, or a target date for draft legislation to reach the House of Assembly. A committee with a mandate and no clock is a promise, not yet a process.

Three Million Dollars, Two Tiers

The money arrived before the committee did anything visible. On August 13, the government confirmed a $3 million investment in AI skills training, financed through an Inter-American Development Bank loan and run with General Assembly, a training provider whose curriculum is built around global employment demand rather than a Bahamas-specific syllabus. Bastian split the funding into two deliberately different tiers: basic AI literacy for the general public, and advanced instruction aimed at Bahamians who want to build AI-driven products rather than only operate the ones other countries export. His stated reasoning was pointed. "If we're going to just start building up all kind of tools and training people to use tools, we're missing a golden opportunity," he told Tribune242, adding that the country needed to "expand our mindset from a degree-based economy" toward one that rewards demonstrated skill regardless of where it was learned.

Trust as the Pacing Mechanism

The day after the funding announcement, Bastian returned to the same theme in different language, telling Tribune242 that public trust, not the underlying technology, would set the speed at which the Bahamas could responsibly adopt AI. Three weeks later, on September 4, he named that idea a National AI Literacy Initiative, describing a programme meant to meet Bahamians at every level, from a first encounter with an AI tool through to professional-grade development work. His framing to the Nassau Guardian was direct: the Bahamas needs people who create inside the AI economy, not only people who consume what has already been built elsewhere. Read against the June committee announcement, the sequence is coherent even if untested: build the literacy base first, let the legislation the committee drafts land on a population that already has some working sense of what it is regulating, rather than a population meeting both the technology and the law for the first time in the same year.

Barbados Already Ran the Small Version

Two islands south, a much smaller version of the same bet had already launched. On September 1, the National Organisation of Women and an AI training firm opened a free programme for up to 35 women, built specifically around developing a viable business idea rather than general tool familiarity. NOW president Melissa Savoury-Gittens put the pitch in plainly economic terms: AI, used well, lets a small entrepreneur do the work she would otherwise pay someone else to do, and that matters most for the businesses with the least room to absorb extra cost. The programme grew out of an earlier NOW initiative that taught young people digital marketing and social media skills, redirected here toward women building something of their own. Thirty-five participants will not move a national economy. What the pilot tests, cheaply and quickly, is whether outcome-specific training beats general literacy training at getting someone from curiosity to a working business, a question the Bahamas' much larger programme has not yet designed a way to answer for itself.

What the Region's Own Numbers Say About Sequencing

CAIA has one dataset that speaks directly to whether training investment changes outcomes rather than just attendance, and it did not come from a government programme. A StarApple AI study of Caribbean organisations that put their boards through structured AI training found governance frameworks stood up in six months rather than the eleven to fifteen months typical without it, organisation-wide literacy scores rose from 2.0 to 3.7, and vendor costs fell by more than 70 percent. That result held at institutional scale, inside organisations that had already chosen to run the programme and had a board willing to sit through it, which is a much narrower and more motivated population than an entire country. Set beside it, the region's adoption baseline is sobering: CAIA's own reporting on the CAITF Final Report found only 13 percent of Caribbean adults use generative AI, against 16.3 percent worldwide, even as regional startups report adoption above 85 percent. The gap between institutions that choose training and a general population that has not been asked yet is exactly the gap the Bahamas' national rollout now has to close.

The Dominican Republic supplies an uncomfortable counterpoint, not a validation. The country's National Statistics Office reported, in a bulletin covered by Dominican Today on September 5, that 35.0 percent of formal Dominican companies with internet access already use AI, rising to 52.0 percent among firms with 250 or more employees, drawn from the 2025 National Survey of Economic Activity. None of that adoption ran through a government literacy programme. It happened because larger, better-connected firms had the incentive and the capacity to adopt on their own, which raises a question the Bahamas has not yet answered in public: is a national literacy fund accelerating adoption that market pressure was already going to produce, or is it reaching the smaller businesses and individual workers that market pressure, left alone, tends to pass over. Those are different jobs, and a single training budget cannot be presumed to do both.

Where the Sequencing Could Break

The risk in training-first sequencing is not that it fails outright. It is that the two tracks drift apart on their own timelines. The National Artificial Intelligence Committee has no published reporting date, while the training investment already has a contracted provider and, per Bastian, will announce public enrolment opportunities in short order. If the committee takes eighteen months to produce legislation while General Assembly-trained Bahamians are already building AI products commercially, the country will spend that gap governing new capability with old law, the exact condition Bastian's own August remarks warned against. The Caribbean AI Association has watched a related version of this problem play out at the regional level, where an endorsed policy roadmap and separate national committees have moved on different clocks without anyone assigned to reconcile them. A national committee with a mandate but no deadline is how that pattern starts.

What CAIA Is Watching

Whether the National Artificial Intelligence Committee publishes a consultation schedule and a reporting date, rather than working toward Cabinet on a timeline nobody outside government can see. Whether the Bahamas releases enrolment and completion numbers for the General Assembly tiers, the same disclosure test CAIA has applied to the Genius Project's own youth programme, instead of stopping at the announcement of the investment. And whether Barbados funds the NOW pilot to run again at ten times its size if the first cohort's businesses actually launch, since thirty-five participants prove a method can work, not that it works at the scale a country the size of the Bahamas needs.

Related Reading Across the Caribbean AI Network

• StarApple AI, the Caribbean's first AI company, on training outcomes that hold up under measurement
• Adrian Dunkley, on building AI capacity inside Caribbean institutions
• Caribbean AI Risk Management Council, on what a national AI governance framework needs to hold up
• AI Barbados, tracking the island's own AI adoption and workforce programmes
• CAIA: the StarApple AI board-training study in full
• CAIA: the Bahamas' earlier bid to write the region's first national AI strategy