Top 5 Jobs in Government That Are Most at Risk from AI in Argentina - And How to Adapt

Argentina's public sector is at a crossroads: deep technical talent and success stories - from Mercado Libre's machine‑learning systems that filter out most fraudulent listings to Buenos Aires pilots that cut stops by 14% - sit alongside under‑deployment of AI and real labour risks, especially for routine clerical, call‑centre and platform roles; the World Bank estimates 26–38% of jobs in LAC are exposed to generative AI while 2–5% face full automation, so whether exposure becomes augmentation or displacement depends on policy and skills investments (World Bank report: Generative AI and Jobs in Latin America and the Caribbean).

Local analysis also flags union concerns about precarious platform work and the need for reskilling, and PANTA's country study maps both Argentina's research strengths and adoption gaps (PANTA country study: Argentina's AI landscape).

Practical, job‑focused training can turn risk into opportunity - see the AI Essentials for Work bootcamp for concrete upskilling pathways to help public servants work with AI, not be replaced by it (AI Essentials for Work bootcamp syllabus (Nucamp)).

Methodology: How We Chose the Top 5 Roles - The selection focused on where generative AI is most likely to replace predictable, routine task bundles rather than bespoke professional judgment: tasks that are high‑volume, rule‑based, and public‑facing (data entry, scripted communication, simple determinations), drawing on Deloitte's three‑criteria framework for assigning work to generative tools (Deloitte: Generative AI for government work tasks).

Next, evidence from public‑sector case studies and workflow automation reviews (FlowForma, Appian, Trinus) guided assessments of technical readiness and common deployment patterns - where low‑code platforms, RPA and copilots already streamline approvals and document routing, automation gains traction quickly (FlowForma guide to government workflow automation).

Finally, worker‑impact research from the Roosevelt Institute ensured risk filters were applied: high exposure alone didn't equal exclusion if tasks still required oversight, language nuance, or contextual judgement; this kept the list grounded in real harms and adaptation priorities for Argentina's municipal and national agencies (Roosevelt Institute: AI and Government Workers), remembering that an automated denial can be “life‑and‑death” for a beneficiary and therefore needs human‑in‑the‑loop safeguards.

“Failures in AI systems, such as wrongful benefit denials, aren't just inconveniences but can be life-and-death situations for people who rely upon government programs.”

Administrative and data‑entry clerks are among the most exposed public‑sector roles because their day‑to‑day work is high‑volume, rule‑bound and public‑facing - exactly the task bundle that city chatbots can automate.

Buenos Aires' Boti shows how fast that shift can happen: deployed across WhatsApp and other channels, Boti now handles millions of citizen conversations each month and has been credited with cutting operational burden roughly in half while centralizing procedure information for easier retrieval (Boti Azure OpenAI service case study - operational impact in Buenos Aires).

Technical choices that matter for clerks' future work include robust input guardrails, a reasoning retriever that pushed top‑k accuracy toward 98.9%, and local‑language prompt engineering so answers use Rioplatense Spanish correctly - features highlighted in the AWS/Bedrock pilot write‑up (Boti Amazon Bedrock pilot - guardrails and retrieval accuracy).

The “so what” is simple: routine entries and scripted replies can be automated at scale, but the practical win for clerks comes from shifting into exception handling, quality‑assurance of AI outputs, and oversight of sensitive decisions where human judgement still matters.

“Boti was our bridge between the government and the citizens when we needed it most.” - Pedro Alessandri, Undersecretary of Smart City, Government of the City of Buenos Aires

For citizen‑facing teams like ANSES and municipal helplines, AI contact centers offer a clear way to turn crushing call volumes into manageable workflows: virtual agents and smart routing can resolve routine eligibility questions and status updates instantly, cut wait times, and free human agents to handle complex, sensitive cases - benefits laid out in Platform28's overview of call‑center AI and Capacity's roundup of AI advantages for government contact centers (Platform28 overview of call-center AI for government contact centers, Capacity roundup of benefits of AI for government call centers).

But the payoff depends on careful design: retrieval‑augmented systems that ground answers in agency records, multilingual support for Argentina's diverse users, and strict data‑security controls so citizen inputs never leak to external training sets.

Governments have lagged private peers in adoption, so starting with small pilots that demonstrate reduced waits, better routing and stronger knowledge‑sharing across agencies is the safest path forward; when done right, the result can feel like replacing a long, frustrating hold loop with a clear, human‑supervised handoff in minutes, not hours (Route Fifty report on governments lagging in adopting AI contact centers).

AFIP tax preparers and routine audit clerks are squarely in the crosshairs because their work is heavy on repetitive number‑crunching, templated forms and rule‑driven checks - the exact inputs hyperautomation targets; industry forecasts show AI is expanding from RPA into end‑to‑end process automation that can triage, flag anomalies and auto‑fill standardized returns (Kaplan Abogados - recent legal and tax reforms in Argentina 2024–2025, including AFIP resolutions that change reporting flows).

Practical hyperautomation trends make clear that routine reconciliations, withholding calculations and bulk audit triage can be automated at scale (Hyperautomation trends for 2025 - ConnectWise), which means many clerks will shift from data entry to exception handling, evidence review and taxpayer communication.

That shift matters: PwC's 2025 AI Jobs Barometer finds workers who adopt AI skills capture significant wage and productivity premiums, so investment in upskilling - prompting, validation checks, and grounded retrieval - turns risk into a career upgrade (PwC 2025 AI Jobs Barometer - AI skills and labor market impacts).

The “so what” is vivid: instead of filing the same dozen forms every day, clerks could be the human safety net that reviews the one flagged case whose outcome affects benefits or municipal revenue - and that judgement will be in demand.

“If you're signing dozens of contracts a week, it's helpful to automate your contracting process…”

Paralegals and court clerks face one of the clearest AI crossroads in Argentina: many core tasks - document drafting, data extraction and routine enforcement case processing - are highly structured and already show rapid gains when paired with generative tools, meaning day‑to‑day workloads can be dramatically reshaped.

Home‑grown systems such as Prometea have cut processing times from weeks to days (the tool has flagged cases and generated drafts at scale, even producing thousands of rulings in record time with reported high accuracy), while research from Fundar and IALAB highlights enforcement trials as ideal candidates for automation because they are repetitive, template‑driven and ripe for safe integration with human checks; see the Prometea judicial AI case study for concrete results and the Fundar & IALAB AI in judicial proceedings report for practical use cases and safeguards (Prometea judicial AI case study, Fundar & IALAB AI in judicial proceedings report).

The “so what” is immediate: clerks who learn to validate summaries, catch hallucinations, anonymize inputs and run DPIAs will move from routine drafter to essential quality‑assurance agents - preserving due process while speeding access to justice - because Argentina's nascent protocols and court pilots demand human oversight and stronger privacy protections before scaling.

“We, as professionals, are not the main characters anymore. We have become editors.”

Radiology technicians and public‑hospital imaging teams in Argentina are on the frontline of a quiet revolution: AI diagnostics and workflow tools can raise diagnostic standards, speed reporting and extend specialist expertise to provincial hospitals, but only if the health system's data and infrastructure are ready.

AI‑assisted reads - already shown to boost cancer detection and, in some pilots, to promise a complete breast readout in under five minutes - could turn high‑volume screening sessions into rapid, actionable results that benefit underserved patients, while technicians pivot toward image quality control, anomaly triage and human‑in‑the‑loop validation (AI-powered radiology trends and diagnostic advances - DeepHealth).

That shift depends on interoperable patient records, secure APIs and scalable platforms - capabilities Argentina has begun building with a national digital health network that links care centers across 24 provinces and handles surges in transaction volumes (Argentina national digital health network case study - Red Hat) - and on regional strategies to improve data quality, workforce training and equity so AI doesn't simply automate scarcity (AI and health in Latin America: equity and policy analysis - Think Global Health).

“With the increased demand in data queries, logging, and COVID-19 case reporting, Red Hat OpenShift helped us collect, manage, and analyze all of that information to provide real-time numbers.”

The path forward in Argentina is pragmatic: pair social dialogue and strong governance with targeted reskilling so automation becomes an upgrade, not a layoff.

Unions must be consulted regularly and involved in designing training and safeguards (see the union reskilling and social‑dialogue recommendations in the GISWatch Argentina report: GISWatch report - Union perception of AI in Argentina), while governments should pilot human‑in‑the‑loop systems and mandatory AI impact assessments before scaling.

Public–private commitments - like Salesforce's $500M pledge for workforce development and Trailhead reskilling in Argentina - create the funding and platforms to scale learning locally (Salesforce $500M investment in Argentina for AI innovation and workforce development).

Tactically, agencies need HR‑led skills audits, small, measurable pilots with strict data controls, and clear retraining pathways so clerks, call‑centre staff and technicians can move into validation, exception handling and oversight roles - for practical, job‑focused training see the AI Essentials for Work bootcamp (AI Essentials for Work bootcamp - Nucamp registration), which maps prompt‑writing and AI‑at‑work skills to everyday public‑sector tasks; the image to hold in mind: union halls turned into Trailhead classrooms, where workers learn the exact prompts and checks that save a citizen's benefits.

“HR is uniquely positioned to shape how AI influences work, how tasks are performed and how work is valued.”