Panama’s AI Readiness Results Put Human-Centered Governance to the Test

Panama has completed UNESCO’s AI readiness assessment, but the results expose a conflict between national ambition and the institutions needed to support it. The country wants artificial intelligence to strengthen logistics, agriculture, public services, and economic growth. It must first close serious gaps in research, data governance, rural connectivity, and public accountability.

The assessment arrives as Panama moves from consultation into implementation. Its government has created new technology governance bodies, advanced AI legislation, and launched a national strategy with five policy pillars. Those steps give Panama more than a statement of intent. They also create deadlines against which the public can judge progress.

The central question is no longer whether Panama wants responsible AI. It is whether human-centered commitments will shape procurement, automated decisions, infrastructure investments, and access outside Panama City. Other Latin American governments face the same test, but Panama’s logistics role and digital connectivity make the outcome especially consequential.

Panama’s RAM Results Turn Principles Into a Policy Checklist

The assessment gives Panama a measurable governance agenda, not a general endorsement of its AI ambitions.

UNESCO’s Readiness Assessment Methodology, or RAM, evaluates whether a country can develop and govern AI in line with ethical principles. It examines legal, social, economic, educational, scientific, and technical conditions. The process combines quantitative evidence with consultations involving government, academia, companies, and civil society.

Panama’s completed country assessment identifies structural strengths alongside significant weaknesses. The resulting report contains 19 policy recommendations. Its priorities include formalizing the national AI policy, consolidating institutional governance, modernizing data protection, and improving public procurement.

That distinction matters. A readiness assessment does not certify that deployed AI systems are safe, fair, or effective. It identifies the conditions that determine whether institutions can evaluate those systems and respond when harms appear.

UNESCO designed the RAM framework to help governments identify missing laws, skills, infrastructure, and oversight capacity. It operates at the national level rather than testing individual models. Panama still needs project-level reviews before agencies deploy AI in sensitive settings.

The assessment found public interest in adoption. A 2025 SENACYT survey reported that 69.4 percent of respondents viewed AI implementation favorably. That support gives policymakers room to act, but it does not remove the need for safeguards.

The consultation process extended beyond a single event in the capital. Four sessions in three cities brought together 106 participants from 45 institutions. Discussions in Chitré and David showed strong interest in practical agricultural applications.

Those regional sessions also produced a proposal described as a Technical Governance Trilogy. It combines ethical standards for public procurement, independent audits for algorithmic systems, and governed open data. Each part addresses a different failure point.

Procurement rules can define acceptable uses before agencies purchase a system. Independent audits can test whether vendors meet those requirements. Governed data access can support development without treating every public dataset as risk-free material.

The proposal remains a policy direction rather than evidence of enforcement. Panama must still define which systems require review, who performs audits, and what happens when a system fails. Without those details, ethical requirements can become contractual language with little operational effect.

This is the first major tension in the RAM results. Panama has converted broad principles into named priorities, but many priorities still lack binding procedures. The next stage requires budgets, responsible agencies, deadlines, and public reporting.

Panama’s AI Strategy Has Institutions, but Capacity Remains Thin

Panama has built an institutional framework faster than it has built the technical workforce needed to operate it.

The country has taken several formal steps during 2026. Bill No. 588, which proposes adaptive AI governance and regulatory experimentation, passed its first legislative debate in April. It had not completed the full legislative process described in UNESCO’s assessment.

Executive Decree 36 followed in May. It established the National Commission for Critical and Emerging Technologies and an AI subcommission chaired by SENACYT. The government formally installed those bodies in July.

A commission can coordinate ministries, universities, and companies. It cannot substitute for trained auditors, data protection specialists, technical evaluators, or public-sector engineers. Panama’s capacity indicators show why staffing deserves as much attention as institutional design.

UNESCO reports that Panama invests 0.16 percent of gross domestic product in research and development. The corresponding Latin American average is 0.33 percent. The country has approximately 71 active AI researchers, compared with a regional average above 1,220.

Panama also has two granted AI patents, while the cited regional average is 160. It has no local doctoral program dedicated to AI. These indicators do not capture every practitioner or commercial project, but they reveal a narrow domestic research base.

The gap creates an immediate implementation problem. New oversight bodies need people who understand model evaluation, privacy law, cybersecurity, procurement, and sector-specific risks. Competition for those skills extends across government, universities, banks, logistics companies, and international employers.

SENACYT has worked with the Georgia Institute of Technology on the national strategy. The earlier strategy work included an ecosystem assessment, strategic alternatives, and an implementation plan.

External partnerships can accelerate training and provide comparative expertise. They also create a dependency risk if domestic institutions cannot retain skills or independently evaluate recommendations. A national strategy needs local ownership after international advisers leave.

Panama’s strategy organizes the work around five pillars. These cover governance, talent, data and infrastructure, ethics and inclusion, and international cooperation. The structure matches the main weaknesses identified by RAM.

The difficult question is sequencing. Panama cannot responsibly expand automated public services without data rules and evaluation capacity. It also cannot build practical expertise without allowing agencies and researchers to test useful systems.

Regulatory experimentation offers one possible bridge. A controlled program can limit deployment scope, require monitoring, and establish exit conditions. Yet a regulatory sandbox is only credible when regulators can reject unsafe projects and publish meaningful findings.

The government therefore faces a capacity race. It must develop oversight skills while commercial and public-sector adoption continues. Pausing every AI project is unrealistic, but deploying systems before controls exist transfers risk to citizens.

The strongest near-term response would connect each priority to a workforce plan. Procurement reform requires trained buyers. Algorithmic audits require qualified reviewers. Data governance requires accountable stewards within every participating agency.

Until those roles are funded and filled, institutional progress will remain uneven. Panama has created the table where AI decisions should be coordinated. It still needs enough qualified people around that table.

Human-Centered AI Collides With Panama’s Digital Divide

A human-centered strategy will fail if rural communities and Indigenous groups remain data subjects rather than decision makers.

Panama’s national averages can conceal sharp differences in access. UNESCO reports internet access in 80 percent of urban households and 46 percent of rural households. That 34-point gap affects who can use AI services, contribute data, and challenge automated outcomes.

An online public service does not become inclusive because it uses a conversational interface. Residents still need connectivity, suitable devices, accessible language, and a non-digital alternative. Agencies must account for those constraints before using automation to replace existing channels.

The linguistic challenge is equally significant. Panama recognizes seven Indigenous languages, and UNESCO classifies all seven as low-resource languages for natural language processing. No digital corpora were available for them when the assessment was prepared.

Low-resource means developers lack enough representative digital text or speech to train and evaluate reliable language systems. A model can therefore perform well in Spanish while failing for citizens who communicate in another language.

Panama joined the Latam-GPT initiative as a strategic partner in February 2026. The collaboration offers a possible route for developing regionally relevant language resources. However, digitizing Indigenous languages raises questions about consent, ownership, access, and commercial reuse.

A corpus is not merely technical fuel. It can include oral histories, culturally sensitive knowledge, personal information, and community-specific expressions. Collecting that material without community governance would contradict a human-centered strategy.

The RAM consultations themselves reveal an early warning. UNESCO notes that comarca representatives were absent. Comarcas are Indigenous administrative regions with distinct political and cultural importance.

That absence does not invalidate the entire assessment. It does mean the process missed voices most affected by language exclusion and rural connectivity gaps. Future consultations need representation before agencies define data-sharing or language technology programs.

Human-centered governance also extends to gender and research participation. Panama has reached gender parity in internet and mobile access according to the cited Digital Gender Gaps Dashboard. Women account for only 39 percent of researchers in the National Research System.

Access to consumer technology is therefore more balanced than participation in its design and oversight. That gap can influence which problems receive funding, how harms are framed, and who gains from public investment.

The global ethics recommendation places human rights, dignity, inclusion, transparency, and environmental protection at the center of AI policy. Panama’s challenge is to convert those principles into choices that redistribute participation.

Public procurement provides a concrete test. A government contract could require accessibility testing, rural service options, language coverage disclosures, and an appeal process. It could also require vendors to document training data and performance differences across groups.

Those requirements would affect product design before deployment. A general ethics statement issued after purchase has much less leverage. Contracts, budgets, and evaluation criteria show whether inclusion is operational or ceremonial.

Agriculture offers another practical setting. Participants outside the capital expressed strong interest in AI for agricultural work. Useful applications might support crop monitoring, weather interpretation, or supply planning.

Yet those tools depend on local data, connectivity, training, and maintenance. A system designed around large farms or stable broadband could deepen existing disparities. Pilots should therefore measure who can use a service, not only whether its predictions appear accurate.

Panama’s inclusion test has a clear standard. Communities must help govern the systems and data that affect them. Consultation after major design decisions will not meet that standard.

The Economic Case Is Large, but the Baseline Is Small

Panama’s projected AI upside is meaningful, yet the gap between potential and current impact should invite discipline rather than celebration.

UNESCO estimates that AI could contribute between $1.776 billion and $3.344 billion annually to Panama’s economy. That range equals roughly 2.1 to 4 percent of gross domestic product. The estimated current impact is only 0.2 to 0.4 percent.

The difference frames AI as a major development opportunity. It also shows that most projected value has not materialized. Forecasts depend on adoption, complementary investment, worker skills, data quality, and institutions capable of managing failure.

Commerce, construction, transport, and storage account for approximately 48 percent of the estimated opportunity. Panama’s logistics position makes transport especially important. The Panama Canal, ports, warehousing, and trade services produce operational data that can support planning and optimization.

AI can assist with forecasting, equipment maintenance, routing, document processing, and risk analysis. Each use requires more than a model. Organizations need reliable data pipelines, domain experts, security controls, and procedures for handling incorrect outputs.

The same dependence applies to public infrastructure. Panama has seven active submarine cables, with two more authorized for 2026. Approximately 65 percent of its electricity comes from renewable sources, according to UNESCO’s profile.

Those assets support Panama’s ambition to become a regional data hub and destination for lower-carbon computing. Connectivity and cleaner electricity can attract infrastructure investment. They do not automatically create domestic research capacity or broadly distributed economic value.

Data centers can consume land, electricity, water, and public incentives while employing relatively small operating teams. Policymakers should distinguish between hosting computing infrastructure and building local capabilities around that infrastructure.

The national strategy needs measurable connections between investment and public benefit. Agreements can address workforce training, research partnerships, energy reporting, local procurement, and service resilience. Without such terms, the hub narrative can outrun its development impact.

Regional comparisons offer useful context. The regional AI index evaluates enabling conditions, research and adoption, and governance across 19 countries. Its 2025 edition classifies ecosystems as pioneers, adopters, or explorers.

Panama sits in the adopter group and improved by 1.47 points from the previous edition, according to UNESCO. That classification recognizes progress without placing Panama among the region’s most mature AI ecosystems.

The result fits the RAM findings. Panama has connectivity, policy momentum, renewable energy, and economic sectors that can use AI. It lacks the research depth, specialized education, and institutional maturity associated with sustained leadership.

Economic forecasts also create political pressure. Once a strategy attaches billions of dollars to AI, officials can feel compelled to accelerate deployments. Vendors can present adoption itself as evidence of modernization.

That pressure makes independent evaluation essential. A public agency should measure whether an AI system reduces processing time, improves service quality, or lowers error rates. It should also measure appeals, exclusions, security incidents, and human workload.

A project that automates a visible task may simply move work elsewhere. Employees might spend more time correcting outputs or explaining decisions. Citizens might face longer disputes when no official accepts responsibility for an automated result.

The most valuable projects will combine clear operational needs with measurable safeguards. Logistics planning, agricultural support, and administrative processing offer opportunities, but each needs a distinct risk assessment.

Panama’s economic opportunity is therefore conditional. The country can gain from AI without treating every deployment as progress. Selective adoption, credible evaluation, and public reporting will produce better evidence than ambitious forecasts alone.

Google News Attention Cannot Replace Independent Accountability

Visibility can amplify Panama’s policy announcement, but only transparent evidence can establish whether its governance model works.

Coverage distributed through Google News can introduce international readers to Panama’s RAM results. It can also compress a complex assessment into a simple story about progress toward human-centered AI.

That framing captures part of the event. Panama has completed a structured national review, created governance bodies, and connected the findings to a national strategy. These are substantive actions.

However, Google News visibility does not verify whether legislation will pass, agencies will follow procurement rules, or audit requirements will become enforceable. Aggregated headlines can make a process milestone look like a completed reform.

The RAM report itself offers a more demanding reading. Nineteen recommendations mean Panama has identified a broad implementation backlog. The country’s legal framework also requires updates for algorithmic profiling, large-scale biometric processing, and synthetic data.

Panama’s Law 81 of 2019 recognizes personal data rights and protections concerning solely automated decisions. UNESCO finds that the law does not fully address several practices associated with modern AI systems.

That gap matters because contemporary services rarely present themselves as a single automated decision. Profiling can influence which cases receive attention, what risks receive scores, and which offers reach a person. Human approval can become superficial when staff routinely accept system recommendations.

Biometric systems raise another level of concern. Errors can affect access, movement, employment, or interactions with authorities. Large databases can also create security risks that persist long after a vendor contract ends.

Synthetic data presents different questions. Artificially generated records can support research when real data is sensitive or scarce. Poorly designed synthetic datasets can still reproduce bias, leak information, or create false confidence about representativeness.

Panama’s proposed independent audit approach needs enough scope to address these cases. An audit should examine data sources, performance, security, accessibility, and the real decision process around a system. A model accuracy score alone is insufficient.

Independence also needs a practical definition. Auditors should not rely entirely on the vendor being evaluated. Agencies need authority to obtain documentation, test systems, and disclose significant findings.

There must also be a remedy when problems appear. Citizens need understandable notices, accessible appeal routes, and a responsible public official. Agencies should retain the ability to suspend systems rather than becoming locked into vendor infrastructure.

The government’s technology agenda connects AI with semiconductor development, science, and national competitiveness. That broader agenda can improve coordination across policy areas.

It can also increase the temptation to measure success through investment announcements and institutional launches. Those metrics matter, but they do not capture whether deployed systems respect rights.

Google News readers should therefore treat the RAM presentation as the beginning of an accountability period. Panama has stated its objectives and documented its gaps. Future decisions can now be compared with that record.

The strongest evidence will come from implementation documents. Procurement standards, audit protocols, agency budgets, consultation records, and incident disclosures can show whether human-centered governance has operational weight.

Three Signals Will Show Whether Panama Can Deliver

The next phase will be decided by enforceable rules, funded capacity, and participation from communities missing in the first assessment.

The first signal is implementation of the 19 RAM recommendations. Panama should identify responsible institutions, deadlines, funding, and public indicators for each priority. A published roadmap would turn a long recommendation list into an accountable program.

Procurement reform deserves early attention because government purchasing can shape systems before they reach citizens. Standard clauses should cover data use, testing, accessibility, audit access, security incidents, appeals, and contract termination.

If agencies adopt those requirements and disclose how they apply them, Panama’s governance claims will become stronger. If implementation remains limited to broad strategy language, the gap between commitment and practice will widen.

The second signal is investment in domestic capacity. Panama needs more researchers, evaluators, public-sector technologists, and data governance specialists. Training totals matter, but retention and institutional placement matter more.

Watch whether universities establish advanced AI study pathways and whether public institutions create permanent technical roles. Partnerships should transfer expertise rather than supply temporary external capacity.

Research spending provides another measurable indicator. The current 0.16 percent share of GDP sits below the cited regional average. A rising share would not guarantee useful research, but flat investment would weaken claims about building a knowledge-based economy.

The third signal is broader participation. Future consultations should include comarca authorities, Indigenous language communities, rural organizations, workers, consumer advocates, and people affected by automated public services.

Their role should extend beyond reviewing a completed plan. Participants need influence over project selection, data governance, risk thresholds, and evaluation. Compensation and accessible participation methods can determine whose input is realistically available.

Language initiatives offer a visible test. Panama can support Indigenous language technology while respecting community consent and control. Clear rules for data ownership, access, reuse, and withdrawal would provide evidence of that commitment.

These three signals reinforce one another. Strong rules require skilled institutions, and skilled institutions need participation to understand real harms. Consultation without authority will not correct weak enforcement.

Panama has valuable assets, including international connectivity, renewable electricity, logistics expertise, and public interest in AI. The RAM results show that those assets coexist with deep institutional and social constraints.

That is why this story is more important than a strategy launch featured on Google News. Panama has published enough evidence to make future performance measurable. It can no longer define progress through ambition alone.

Developers should watch procurement standards because they will shape technical requirements. Enterprise buyers should watch audit and data rules because they can influence vendor expectations across the market. Knowledge workers should watch training and appeal mechanisms because automation will change both responsibilities and accountability.

The next question is concrete: will Panama publish a funded roadmap that connects every major AI deployment to oversight, participation, and measurable public value? That document, followed by visible enforcement, would show that human-centered AI is becoming an operating model rather than a headline.