Digitalisation or dependency? Sri Lanka’s AI dilemma beyond AKD’s binary

A binary “digitalise or perish” narrative must give way to a measured path

1.The presidential frame: Digitalise or perish

President Anura Kumara Disanayake has cast digital transformation as a stark binary: either embrace artificial intelligence (AI) and modernise, or be “pushed to the margins” of the global economy. His appeal identifies inefficiency, stagnation, and weak public services as consequences of past inaction. This rhetoric mirrors global leaders who treat AI as a “pillar” of competitiveness and a survival imperative.

Yet such framing is both motivational and reductionist. It risks presenting digitalisation as a technological silver bullet, obscuring the deeper questions of sovereignty, governance, and inequality that must shape Sri Lanka’s pathway.

1.1 Against the fiction of AI doom

Adam Becker cautions against equating AI with an autonomous “brain” destined to rebel against humanity. The metaphor of AI as a computer like mind is flawed, he argues, and “existential risk” talk distracts from tangible, present day harms1.

Applied to Sri Lanka, Becker’s insight affirms AKD’s focus on practical adoption rather than sci-fi apocalypse. But it also warns against over investment in hype: when policymakers believe AI to be omnipotent, they risk neglecting systemic risks such as bias, inequity, and monopolistic dependency.

2.Real-world risks for Sri Lanka

If digitalisation proceeds without safeguards, the consequences could be profound:

• Bias and discrimination in welfare or education systems.
• Reliability failures, with AI “hallucinations” producing false outputs.
• Opacity, as citizens cannot appeal decisions made by black box systems.
• Data privacy gaps, risking surveillance or foreign capture of personal information.
• Misinformation and manipulation, with deepfakes fuelling polarisation.
• Election misinformation, where false stories, manipulated images, and AI generated videos undermine democratic debate. During the 2024–25 elections, disinformation campaigns spread rapidly across social media, often micro targeted at voters. Without regulation, this erodes trust in the Election Commission, in news media, and in democracy itself2.
• Job displacement, particularly among urban youth in clerical and service sectors.
• Environmental costs, from energy intensive data centres.
• Concentration of power, as digital infrastructure is monopolised by foreign firms.
• Weaponisation of AI, given Sri Lanka’s history of surveillance and conflict.
• Regulatory vacuum, with no ethical AI framework in place.

3.Debt arithmetic and why a blind sprint can backfire

Sri Lanka’s fiscal room to manoeuvre is vanishingly small. Under the IMF program, Government must run primary surpluses and push public debt down from roughly 110% of GDP toward 80–85% by 20323, 4. In such a setting, every rupee funnelled into prestige digital projects crowds out what is urgently needed to cushion disruption: unemployment protection, retraining, and sectoral transition support. But Sri Lanka has no nationwide unemployment insurance yet, and more than half of the poorest citizens were not covered by safety nets even during the crisis—evidence that administrative fixes and “smart targeting” do not magically create buffers where fiscal capacity is thin5.

Guardrails are also late. The Personal Data Protection Act only began full enforcement in March 2025; institutions, capacity and independence still need to mature before large biometric or AI systems are expanded6. Meanwhile, energy intensive data centres and AI infrastructure face tariff volatility: a 15% electricity hike in June 2025 underlines the recurrent cost risk to Government and SMEs alike7.

The policy implication is sequencing, not speed: build legal safeguards, labour protections, and human capital first; target digital spend to high-value domestic problems (health, agriculture, logistics) with open standards and clear exit options. Otherwise digitalisation risks deepening inequality while adding hard recurrent costs to a debt-burdened state.

4. The Indo–Lanka Digital ID case: A sovereignty warning

These risks are not abstract. They are visible today in the Indo–Lanka Digital Identity (SLUDI) project, now under Supreme Court challenge.

In January and June 2025, Cabinet amended the original 2022 MoU with India, allowing New Delhi to select and control the project’s core technical infrastructure, including the biometric database software (MOSIP) and Master System Integrator (MSI). Petitions filed by former Minister Wimal Weerawansa and civil society activists allege that the deal jeopardises Sri Lanka’s sovereignty and exposes citizens’ biometric data to foreign misuse8.

The Supreme Court has issued notice to the President and Cabinet, recognising at least a prima facie concern. This controversy highlights the danger of “modernisation by outsourcing”: in digital governance, whoever controls the database controls the citizen.

5.AI and the call centre mirage

Another overlooked risk is employment. Sri Lanka has marketed itself as a destination for call centres and back office outsourcing. Yet advances in AI voice agents, customer service chatbots, and automated translation threaten to erase this industry.

The paradox is cruel: while Colombo may trumpet digitalisation as a source of new jobs, global AI firms headquartered in California, London, or Bangalore are already automating these very functions. Sri Lanka has little leverage to regulate these giants or shield its workforce. For thousands of young urban workers, the dream of digital employment could become a nightmare of redundancy.

6.Farmers and fisherfolk: The panacea exposed

President Disanayake frequently highlights foreign examples of digitalisation: fishermen abroad using mobile signals to locate schools of fish, or farmers applying digital forecasts to guide crop cycles. Yet these narratives obscure the actual burdens of Sri Lankan farmers and fisherfolk.

For farmers, the crisis is not the absence of a weather app. It is that their crops are destroyed daily by elephants, monkeys, and other pests, without an effective eradication or control program. Structural reforms are equally urgent:

• ensuring fair prices for produce,
• building efficient marketing systems,
• strengthening crop insurance, and
• expanding financing facilities.

Digital tools may play a role, but without addressing these fundamentals, they offer little more than a superficial solution.

For fisherfolk, the problem is not that they cannot locate fish shoals in the ocean. The greater threat is Indian bottom trawling, conducted illegally and destructively in Sri Lankan waters most nights of the week. These industrial vessels devastate marine ecosystems and make it impossible for small-boat fishermen to compete.

Digitalisation cannot protect livelihoods when the seas themselves are stripped bare by industrial exploitation. What is needed is enforcement, diplomacy, and sustainability — not a mobile signal pointing to waters dominated by foreign trawlers.

Adam Becker cautions against equating AI with an autonomous “brain” destined to rebel against humanity. The metaphor of AI as a computer like mind is flawed, he argues, and “existential risk” talk distracts from tangible, present day harms. Applied to Sri Lanka, Becker’s insight affirms AKD’s focus on practical adoption rather than sci-fi apocalypse. But it also warns against over investment in hype: when policymakers believe AI to be omnipotent, they risk neglecting systemic risks such as bias, inequity, and monopolistic dependency

7.The publishing and knowledge trap

Alex Reisner’s reporting shows how AI platforms are hollowing out journalism and creative industries by siphoning content without compensation9. Tools such as Google’s AI Overviews and ChatGPT increasingly substitute for original news outlets, cutting traffic by up to 34%. For small states like Sri Lanka, where journalism is already fragile, this represents not just an industry crisis but a democratic one: without viable publishing, investigative journalism collapses, and sovereignty of information is outsourced to Silicon Valley.

The same dynamic applies to YouTube creators. More than 15 million videos, many instructional or educational, have been scraped into AI datasets10. If Sri Lankan educators, media outlets, or artists upload content today, tomorrow it may fuel AI systems that compete against them, undermining local creative economies.

8.A Sri Lankan pathway

A binary “digitalise or perish” narrative must give way to a measured path:

• Governance first: data protection laws, ethical AI frameworks, and accountability mechanisms.
• Local adaptation: Sinhala/Tamil datasets, agriculture and health-focused AI, not blind imports.
• Human capital: digital literacy for administrators, teachers, and youth.
• Equity lens: labour safeguards and inclusion in digital transformation.
• Sustainability: energy-conscious AI use, aligned with debt and climate realities.
• Knowledge sovereignty: policies to protect local publishers, educators, biometric data, and fisherfolk from unregulated extraction.

9.Conclusion

President Disanayake is right that Sri Lanka cannot afford digital inertia. But a “two roads” metaphor is misleading. Digitalisation is not destiny; it is a political and ethical project.

Becker reminds us AI is not a brain, while Reisner warns that unregulated AI erodes journalism and creativity—the lifeblood of democracy. The Indo–Lanka Digital ID case offers a prima facie reason for vigilance: sovereignty, once outsourced, cannot easily be reclaimed. And for farmers and fisherfolk, digitalisation without addressing material realities is simply irrelevant.

For Sri Lanka, the real challenge is not whether to digitalise, but how to digitalise without mortgaging its sovereignty, its jobs, and its future.

Footnotes:

1Adam Becker, “The Useful Idiots of AI Doomsaying,” The Atlantic, 19 September 2025.

2On digital disinformation and election risks, see global coverage of the 2024–25 Sri Lankan cycle; cf. Becker, ibid.

3Reuters, “Sri Lanka eyes 6% growth in 2026; debt reduction to 80–85% of GDP by 2032,” 15 September 2025.

4International Monetary Fund, Sri Lanka—Third Review under the Extended Fund Facility, 28 February / 3 March 2025.

5World Bank, “3 ways to make social protection work for Sri Lanka’s poor and vulnerable,” 19 June 2023.

6Data Protection Authority of Sri Lanka, “PDPA dates of operation—full enforcement from 18 March 2025.”

7Public Utilities Commission of Sri Lanka (PUCSL), tariff revision decision; see also News 1st report on 15% increase, 11 June 2025.

8“Another petition filed challenging Indo–Lanka Digital Identity deal,” Daily Mirror, 22 August 2025; “Sri Lanka’s Supreme Court issues notice to govt on India-funded digital ID project,” PTI/Colombo, 27 August 2025.

9Alex Reisner, “The End of Publishing as We Know It,” The Atlantic, 25 June 2025.

10Alex Reisner, “AI Is Coming for YouTube Creators,” The Atlantic, 10 September 2025.

Abstract:

 Sri Lanka must digitise, but not at the expense of sovereignty, fiscal stability, and social cohesion. This article argues that a debt-burdened state cannot sprint into prestige technologies without first building guardrails: enforceable data protection, labour protections, and sector-specific human capital. Drawing on the Supreme Court challenge to the Indo–Lanka Digital Identity project, the analysis shows how “modernisation by outsourcing” can compromise control over citizens’ data. It also highlights near-term risks to employment (call centres), weak safety nets, volatile energy costs for data infrastructure, and the erosion of local journalism and education through AI intermediaries. The path forward is sequencing: governance and capability first, targeted digital investment second—so that digitalisation serves Sri Lanka’s people rather than indebting or displacing them.

(The writer is a political economist and researcher who writes on constitutional law, governance, and political economy. Contact: [email protected].)