By Kirk Rampersad
A few weeks ago, I watched The AI Doc: Or How I Became an Apocaloptimist on Netflix. The documentary, directed by Daniel Roher and Charlie Tyrell, follows a father-to-be as he tries to understand the world his child will inherit. Its questions about artificial intelligence stayed with me because they reach beyond technology into parenting, education, work and trust.
One remark was particularly unsettling. Sam Altman predicted that our children would never surpass AI from a “raw IQ perspective”.
Although he expressed optimism about parenthood, I found the implications harder to dismiss.
Beyond the documentary, developments once confined to science fiction are making the future feel remarkably close. On October 15, Tesla demonstrated a Roadster prototype capable of brief, thruster-powered flight, a striking engineering achievement, although not evidence that predictions about AI will materialise.
Elon Musk has separately predicted that AI could surpass any individual’s intelligence as early as 2026 and eventually exceed humanity’s collective intelligence. He has also forecast superhuman performance in digital tasks by the end of 2027, and further suggested that AI could outperform humans across all fields by 2028. These are predictions, not certainties. Nevertheless, they sharpen the concern that stayed with me after the film: in our eagerness to build machines that can do almost anything, are we preparing our children to think for themselves or to accept a future in which thinking is increasingly done for them?
However, machines becoming more capable does not establish that children will have lower IQs. The more immediate concern is whether we are creating conditions in which people no longer practise the thinking that develops their own abilities.
For T&T, that distinction matters. Our future cannot rest on access to intelligent software without the human judgment to challenge it.
The convenience trap
The film raises the prospect of powerful systems escaping human control. Such risks deserve scrutiny, not acceptance as inevitable. But a quieter danger is already emerging: the voluntary surrender of our judgement.
Consider a child submitting an excellent assignment that they cannot explain, or an employee presenting a recommendation without understanding its assumptions. The output looks successful. The learning may never have happened.
A 2025 study in the Proceedings of the National Academy of Sciences examined nearly 1,000 students at one Turkish high school. Those using a basic AI tutor improved during assisted practice but scored 17 per cent lower than the control group on exams without AI. Learning safeguards largely prevented that harm.
This short-term mathematics study does not establish permanent damage or a decline in childhood IQ. Its warning is important: assistance that improves immediate performance can undermine independent learning.
Microsoft Research and collaborators also surveyed 319 knowledge workers in 2025. Greater confidence in AI was associated with less reported critical-thinking effort. This does not prove lasting decline, but challenges the assumption that polished output reflects deeper thought.
The risk is a generation that can obtain almost any answer but has less practice deciding whether that answer deserves belief. A country fluent in prompting but unable to question the answers risks dependence, particularly when systems influence livelihoods, education and opportunity.
Beyond task completion, what did the person understand, question and retain?
The performance of expertise
The same concern follows us onto LinkedIn and other social platforms.
In June, LinkedIn described a rise in low-effort, AI-generated content lacking original perspective. It announced measures against generic content and automated interactions while accepting AI as a writing aid.
Refining language is not inherently dishonest. Fabricating experience, claiming unsubstantiated expertise or presenting arguments one cannot defend is different.
A polished paragraph proves neither the use of AI nor competence. We should resist automatic suspicion and admiration alike.
What concerns me is professional visibility becoming detached from substance. Generating the language of leadership is not exercising leadership, however convincing the presentation or confident the person behind it may be.
Employers and clients need better tests than fluency and frequent posting: can this person explain their recommendations, answer difficult questions and deliver results?
T&T’s competitiveness depends on solving real problems, not merely creating impressive professional profiles.
When fakery reaches our pockets
The stakes extend beyond professional vanity.
In March 2026, the Trinidad and Tobago Securities and Exchange Commission warned about a fraudulent investment scheme that used AI-generated videos, fabricated newspaper articles and altered images to fabricate endorsements.
This was technology making deception look credible.
A familiar face, an authoritative voice, or a newspaper-style layout cannot substitute for verification. Before sharing sensational claims or acting on investment promotions, we should check independently through established channels. A family WhatsApp group should not become a fraudster’s distribution network.
Beyond scams, the International Labour Organisation estimates that one in four workers globally is in an occupation with some exposure to generative AI. It considers transformation more likely than wholesale replacement; exposure does not forecast job losses.
Businesses must also consider what happens when entry-level tasks are automated before younger workers learn from them. Saving time today could weaken tomorrow’s pipeline of experienced professionals. Training cannot be somebody else’s responsibility.
A future worth choosing
None of this requires rejecting AI.
A 2025 Quarterly Journal of Economics study of 5,172 customer support agents found that AI assistance increased the number of issues resolved per hour by 15 per cent on average, with substantial benefits for less-experienced workers. That is meaningful evidence of productivity gains in a specific setting, not a guarantee for every occupation.
For a small business in T&T, time saved on routine administration could support better customer relationships and market development. Properly designed tools could also help people learn rather than merely complete assignments.
The choice is therefore about design, discipline and accountability.
T&T has already participated in an AI-readiness validation workshop involving the Ministry of Public Administration and Artificial Intelligence, UNESCO and the University of the West Indies. Held in February 2026, the event addressed the foundations of ethical, inclusive AI governance.
That national conversation belongs in everyday settings too.
Schools should require students to explain their reasoning, show drafts and demonstrate selected skills without assistance. Teachers need support in designing these assessments and students need clear rules on permitted AI use. The goal should be understanding that survives when the screen is closed.
Parents should set aside time for reading, conversation, play and problem-solving without instant answers. Asking a child, “How did you work that out?” may become more important than admiring a perfect submission.
Employers should establish approved uses, protect confidential information and retain named human responsibility for consequential decisions. Recruitment and promotion should test demonstrated ability, not merely polished applications or online visibility.
Staff should be able to disclose legitimate AI assistance without feeling driven to hide it.
Professional bodies and community organisations should help people recognise fabricated content and verify claims. Access matters too: responsible adoption must include those who cannot afford premium tools or specialist training.
At the level of powerful AI systems, independent testing, transparent reporting and meaningful oversight should complement company assurances. A society cannot responsibly delegate both the creation of risk and the definition of acceptable risk to those selling the technology.
We also need to judge success differently. Counting subscriptions, generated reports or minutes saved will tell us little about whether people are becoming more capable. Schools and workplaces should also examine accuracy, independent understanding and the ability to recognise when an answer is wrong. Efficiency should strengthen human capability, not quietly replace the practice through which it develops.
The documentary’s title holds anxiety and optimism together. That seems an appropriate starting point, provided optimism is attached to responsibility rather than wishful thinking.
The question for T&T is not whether our children will outperform machines at everything. They will need to grow into people who can recognise error, exercise judgement, care for others and take responsibility for their choices.
Let the machines become more capable. Our responsibility, as parents, professionals and citizens, is to ensure that the people of T&T do too.
Kirk Rampersad can be contacted via email: kirkram@hotmail.com or at linkedin.com/in/kirk-rampersad-mba-5ab579268
