– Raju Bhattarai
Artificial intelligence is shifting from pilot to production across every major sector. In 2024–2025, AI moved beyond hype cycles into hard economics (capex, opex, productivity), governance (national strategies, education policy), and everyday use (from classrooms to creator tools). Below is a concise, data-grounded synthesis you can run on your site.
1) The Global AI Economy: Investment, Spending, and Diffusion
Private & corporate investment. In 2024, U.S. private AI investment reached $109.1B, nearly 12× China and 24× the U.K.; total corporate AI investment (private + M&A + grants) hit $252.3B. Generative AI alone drew $33.9B globally. Business usage rose fast: 78% of organizations reported using AI in 2024 (up from 55% in 2023).
Enterprise spend outlook. IDC projects enterprises will spend $307B on AI in 2025, rising toward $632B by 2028; AI infrastructure spending is also surging through 2028 as cloud-based AI servers dominate deployments.
Big Tech capex arms race. Analyses now model hyperscaler AI capex surpassing $350B in 2025 and approaching $400B+ in 2026, driven by data centers, accelerators, and networking—illustrating the “compute as moat” dynamic.
Capital flows are concentrating in model development and infra, while adoption broadens across mainstream enterprises—an indication that AI is embedding into the global production function rather than remaining a niche toolset.
2) What AI Does Well (and Where It Hurts)
Advantages (evidence-based)
Productivity & decision support: Enterprises cite efficiency gains in customer ops, analytics, coding assistance, and content workflows—consistent with adoption jumps measured in the AI Index 2025.
Healthcare & science: Image analysis, triage, and drug-discovery pipelines increasingly use AI; regulators are moving from sandbox to scaled evaluation frameworks (see UNESCO’s push for sector guidelines in education/research, mirrored by health agencies globally).
Accessibility & inclusion: Speech interfaces, captioning, translation, and adaptive interfaces lower participation barriers across languages and abilities—again reflected in enterprise adoption data.
Disadvantages / Risks
Labour market transition: The World Economic Forum’s 2025 outlook anticipates both job creation and displacement through 2030; recent summaries cite ~170M jobs created vs. ~92M displaced globally if reskilling keeps pace. Early U.S. evidence shows entry-level, AI-exposed roles for young workers have already contracted, while mid-career roles see net gains—indicating near-term disruption at the bottom of the ladder.
Bias, privacy, and security: OECD finds AI exposure reshapes task profiles and skill demands (management, process, and communication skills rise), while risks from data misuse and biased outputs persist.
Cost concentration & fragility: The hyperscaler capex race raises barriers to entry; delivery costs and reliability constraints may limit near-term ROI for some use cases despite large efficiency wins.
3) Children and Young People: Actual Engagement Patterns
Usage is widespread and rising
In the U.K., 50% of 8–17s used AI tools in 2025 (up from 46% in 2024). Among online teens 13–17, ~79% report using generative AI tools.
Literacy Trust (2024) shows ~77% of 13–18-year-olds used gen-AI, up from 37% in 2023, reflecting exceptionally rapid diffusion in one year.
National Literacy Trust
U.S. surveys indicate ~70% of teens have used a gen-AI tool; usage frequency varies (few are daily users).
What they’re doing with it
Homework help, coding snippets, language practice, summarization, and creative ideation dominate. Educators simultaneously report gains in productivity and fresh concerns about originality and over-reliance—mirroring UNESCO’s call for competency frameworks for students and teachers.
Safety & wellbeing concerns
Investigations and news coverage highlight risks from AI companions in teen contexts (emotional dependency, mishandled crisis content). Common Sense Media reporting suggests 31% of teens found AI chats as satisfying as real friendships, and one-third discussed serious issues with AI rather than people—intensifying calls for youth-appropriate design and stronger guardrails on mainstream platforms.
4) Policy, Standards, and School-System Readiness
UNESCO guidance (2023–2025): Countries are encouraged to adopt AI competency frameworks and gen-AI guidance for education and research—prioritizing safe, human-centered use, teacher training, assessment integrity, and data protection.
OECD skills evidence: Across 10 OECD countries, jobs exposed to AI are shifting toward management, process, communication, and tech-complementary skills even when specialized AI skills aren’t required—so curricula must couple foundational literacies (math, writing, statistics) with AI-aware workflows.
Build AI literacy across grades (prompts, verification, bias, IP, privacy).
Integrate assessment redesign (process logs, oral defense, project portfolios).
Provide teacher PD on model limits and classroom integration aligned with UNESCO’s frameworks.
5) The Labour Market: What to Prepare Young People For
Signals from 2025 reports
Employers rank AI & information processing as the most transformative technology this decade; “broadening digital access” is the cross-cutting enabler.
World Economic Forum
Scenario bands suggest net job creation if reskilling is scaled, but short-run displacement of routine cognitive work (support, entry-level analysis, basic coding) is already visible in microdata on young workers.
Skill clusters with durability
AI-complementary cognitive: statistics, data literacy, critical reading, structured writing.
Human-centric: interpersonal communication, ethics, compliance, client discovery, leadership.
Systems: product thinking, process design/QA, governance, prompt + tool orchestration.
These align with OECD findings on shifting demands in AI-exposed roles.
6) 2025–2030 Outlook: Three plausible scenarios (with leading indicators)
Augmented Productivity, Managed Risk
Enterprises scale copilots; education adopts guardrailed AI tutors; national frameworks mature. Productivity gains compound; net job growth materializes as re-training catches up. Watch: enterprise AI spend tracking IDC curves; school systems adopting UNESCO-aligned frameworks.
Concentration & Bottlenecks
Compute/capex concentration plus high delivery costs stall smaller players; ROI narrows to a few scalable use cases; skills mismatches widen youth unemployment in AI-exposed entry roles. Watch: hyperscaler capex concentration, persistent entry-level job declines in exposed occupations.
Safety & Trust Backlash
High-profile safety failures in youth contexts (companion chatbots, harmful prompts) trigger stricter platform rules and education mandates; adoption slows in schools until certified, age-appropriate models dominate. Watch: policy responses aligned to UNESCO guidance and national regulators.
7) Action Playbook (for schools, parents, and youth programs)
Adopt an AI Use Policy anchored to UNESCO guidance (consent, data minimization, bias checks, age-appropriate guardrails).
Teach Verification: require students to show reasoning, sources, and iteration history for AI-assisted work.
Portfolio-first Assessment to reduce “copy-paste” incentives and reward process mastery.
Wellbeing Protocols: prohibit AI companions for at-risk students; route sensitive queries to human support; ensure crisis-response escalation works as designed.
Skill Pathways: project modules that pair domain learning (science, business, media) with AI tools and reflection on limits (hallucinations, bias, privacy).
Youth Work Transitions: emphasize internships and apprenticeships that cultivate the AI-complementary skills OECD identifies.
