Andrea Schiona

Posted on Aug 16

AI and the Future of Work: Risks, Numbers, and New Careers — AI Breakfast Special

#ai #futureofwork #career #technology

AI and the Future of Work: Risks, Numbers, and New Careers

AI Breakfast Special — What the data really says about automation, job losses, and the roles being created right now.

The public debate on AI and work is polarized between two narratives. On one side, the jobs apocalypse: robots and algorithms replacing millions of people. On the other, pure technological optimism: AI as a purely job-creating force. The data from major international organizations tells a more nuanced — and more urgent — story. This is not mass replacement, but a uneven transformation: job contents are changing, new roles are emerging, and inequalities are widening between those who adapt and those who are left behind.

This article breaks down the numbers, the most exposed sectors, the new careers already visible in the market, and the policies needed to steer the transition, with a specific focus on Italy and Europe.

The Numbers: Job Losses vs New Jobs

WEF Future of Jobs Report 2025/2026

The World Economic Forum estimates that by 2030, automation and AI adoption could make approximately 85 million jobs obsolete globally. But the same report predicts the creation of approximately 170 million new jobs, for a net positive estimated at +85 million. The catch: this figure is aggregated. The geographic, sectoral, and skill-level distribution is highly uneven.

McKinsey Global Institute — Agents, robots, and us (Nov 2025)

McKinsey corrects the doomsday narrative: this is not sudden mass layoffs, but a prolonged restructuring of work content. About 40% of tasks currently performed by workers could be automated by 2030, but only a minority of these tasks correspond to entire professions destined to disappear. More common is the case of roles that become hybrid: an administrative worker moving from data entry to AI workflow supervision, a lawyer integrating automated document review tools.

OECD Employment Outlook 2024

The OECD stresses that impact varies drastically by sector and skill level. Workers with low qualifications and repetitive tasks are the most exposed, while professions requiring high human interaction, creativity, and complex problem-solving show structural resilience. The organization warns, however, that without reskilling policies the risk is increased wage and territorial inequality.

IMF World Economic Outlook — AI and work (Apr 2025)

The International Monetary Fund estimates that AI could contribute to a 2-3% increase in global productivity in the coming years, but with asymmetric distributional effects. In countries with low digitalization — including Italy — delays in adoption and training widen the gap with more advanced economies.

The Most Exposed Sectors: Who Is at Risk

Not all jobs are equally at risk. Automation hits hardest activities with these characteristics:

Structured repetition: data entry, filing, classification
Explicit, predictable rules: standardized administrative procedures
Simple textual interaction: FAQ responses, first-line customer support, basic ticketing

The sectors where impact will be fastest and deepest:

Administration and offices: basic accounting, practice management, invoicing, document logistics
Customer service and call centers: chatbots and virtual assistants already handle first-line support today
Traditional manufacturing: collaborative robots and predictive AI reduce the need for operators on standardized lines
Retail and entry-level logistics: warehousing, deliveries, automatic checkout, inventory
Transport: autonomous driving for long-haul routes and internal logistics

Conversely, more resilient are professions requiring:

Emotional intelligence and human connection: nurses, teachers, therapists, social workers
Creativity and strategy: designers, architects, researchers, managers
Unstructured physical work: plumbers, electricians, healthcare operators
Complex ethical and legal decisions: judges, lawyers, auditors

New Roles Born with AI

The other side of the coin is the rapid emergence of professions that didn’t exist ten years ago. These roles are not marginal: they require hybrid skills, are well-paid, and in strong growth.

AI Trainer and Prompt Engineer

These professionals train, calibrate, and optimize AI models for specific tasks. They don’t just write prompts: they design fine-tuning datasets, define behavioral policies, evaluate response quality, and manage alignment between model and business objectives. In Italy, demand still lags behind qualified supply, but is growing in finance, retail, and public administration.

AI Governance Officer and Compliance Specialist

With the EU AI Act in force since 2024, high-risk AI systems must comply with transparency, documentation, and human control obligations. This has created a new professional pipeline: the governance officer defines internal policies, the compliance specialist verifies legal conformity, the risk assessor maps algorithmic risks. By 2027, every company with more than 250 employees will likely need at least one dedicated figure.

MLOps Engineer

Between data science and operations, the MLOps engineer builds and maintains the pipelines that bring AI models into production. Manages model versioning, performance monitoring, continuous updates, and integration with enterprise systems. It’s one of the most sought-after tech roles, with 70% growth in the last two years.

Data Curator and AI Content Specialist

AI models feed on data. The data curator selects, cleans, labels, and organizes high-quality datasets. It’s a hybrid figure between data scientist and archivist, increasingly important in healthcare, finance, and media. The AI content specialist, meanwhile, produces and reviews AI-generated content, ensuring accuracy, brand tone, and regulatory compliance.

Red Team Specialist and AI Safety Researcher

When AI systems become critical — in cybersecurity, medicine, mobility — professionals are needed to test them systematically. The red team specialist looks for vulnerabilities, bias, unexpected behaviors. The AI safety researcher works on alignment theory and practice, in contexts ranging from labs to regulators.

Emerging Hybrid Roles

Beyond specialized roles, hybrid professions are growing: a doctor using AI-assisted diagnostic tools, a lawyer integrating predictive jurisprudential analysis, a journalist verifying algorithm-generated content, a project manager orchestrating human teams and AI agents. The trend is not replacement, but integration.

The Italy Case: Between Delay and Opportunity

Italy starts from a disadvantaged position in the AI adoption race, but the PNRR offers unprecedented resources to close the gap.

Data and Analysis

According to Banca d’Italia in the 2024 Annual Report, adoption of advanced digital technologies in Italian productive fabric is still limited: less than 20% of enterprises use AI tools structurally. Unioncamere Excelsior 2025 confirms that among workers at risk of transformation are mainly service-sector employees and specialized workers in routine processes.

CEDEFOP Skill Anticipation 2025 estimates that in Italy between 3 and 5 million workers could see their roles deeply transformed by 2030. The risk is not mass unemployment, but the “impiegificio”: underqualified positions, stagnant wages, working poor. To avoid it, a large-scale reskilling strategy is needed.

The PNRR as a Lever

The National Recovery and Resilience Plan allocates significant resources to digital training, AI research, and industrial transition. The problem is implementation speed: small and medium enterprises struggle to access calls, training paths are often misaligned with market needs. An ecosystem is needed where universities, research centers, companies, and regions work together.

Italy’s Cultural Gap

Beyond technical skills, there’s a problem of trust and understanding of AI. In Italy, around 60% of workers declare they don’t know AI tools, and only 15% use them regularly in their work. This generational and cultural divide is the real bottleneck: it’s not just about courses, but a change of mindset.

Europe and the EU AI Act: Rules That Create Jobs

Europe has chosen regulation over laissez-faire. The EU AI Act classifies AI systems by risk and imposes increasing obligations: transparency, documentation, human control, periodic audits.

New Regulatory Professions

This regulation is generating demand for new professional figures:

AI Compliance Officer: verifies systems comply with regulations
AI Auditor: conducts independent evaluations of high-risk models
Ethics Officer: oversees ethical and social impact of algorithms
Human-in-the-Loop Designer: designs interfaces and processes that maintain human control

By 2030, around 500,000 new jobs related to AI governance are estimated to emerge in Europe. In Italy, with well-oriented PNRR and active policies, we can capture a significant share of this demand.

Global Competition

Europe is not alone in the AI race. The United States and China invest far more in research and adoption. The risk is that Italy and Europe become consumers of others’ technologies, without developing their own supply chain. The alternative is investing in applied research, startups, and training, transforming regulation from a cost to a competitive advantage: trust in the European AI system can become an asset.

The Real Risk: Uneven Transition

AI is not a uniform event. It hits some sectors first, then others; benefits some categories, penalizes others; helps some regions, leaves others behind. The greatest danger is not total unemployment, but polarization.

Territorial Inequality

Northern Italy, with a more industrialized productive structure and more dynamic SMEs, is better positioned to adopt AI and attract investment. The South, with greater exposure to traditional sectors and less digital infrastructure, risks suffering the transition without benefiting. Without convergence policies, AI could widen the North-South gap instead of reducing it.

Generational and Gender Inequality

Workers over 50 with limited digital skills are the most vulnerable. Women, overrepresented in administrative and service sectors at risk of automation, could suffer a disproportionate impact. Meanwhile, new AI roles today are predominantly male and young: active inclusion policies are needed.

The “Impiegificio” Risk

If the transition is not governed, the danger is not unemployment, but mass dequalification: workers expelled from traditional roles ending up in precarious, low-skill, low-pay positions. AI then becomes a tool of wage compression, not growth. History teaches that technological revolutions create wealth, but only when accompanied by investment in human capital.

What to Do: Individual and Systemic Strategy

For Workers

Continuous training: a degree is not enough. Every 3-5 years, update your skills. Platforms like Coursera, Udacity, and PNRR initiatives offer recognized paths.
Hybrid skills: combine your domain expertise with AI knowledge. An accountant learning AI for audit, a marketer learning AI-assisted content generation, a teacher integrating AI tools into teaching.
Develop “soft AI literacy”: you don’t need to be a data scientist, but you need to evaluate output, recognize bias, and manage human-machine interaction.

For Companies

Gradual and participatory adoption: AI is not installed, it is integrated. Involving workers in designing new workflows reduces resistance to change and increases transformation value.
Internal reskilling plans: companies that invest in requalifying personnel have higher AI adoption rates and lower turnover.
Governance before technology: define policies, responsibilities, and controls before implementing high-risk AI systems.

For Public Policy

Recognized micro-credentials: fast, flexible, interoperable AI skill certification systems across regions and countries.
Lifelong learning accounts: a personal fund for every worker, financed by companies and the State, for continuous training.
Active labor policies: transition support, income support during training, incentives for hiring in strategic sectors.
Research and innovation: increase public investment in AI applied to traditional Italian sectors — precision agriculture, tourism, manufacturing, healthcare.

The Bottom Line: Govern, Don’t Defend

Artificial intelligence does not destroy work uniformly. It transforms it. The real risk is the uneven transition: those who update their skills become more productive and better paid, those who don’t risk exclusion.

The goal is not to defend against automation, but to govern it. This requires three things: widespread skills, clear rules, and bold investment. Italy has the opportunity to use AI to close its structural gaps, but only if it faces the transition as a collective project, not as a sum of individual choices.

The future of work is not human against machine. It is human with the machine.

Sources

McKinsey Global Institute — Agents, robots, and us (Nov 2025)
WEF Future of Jobs Report 2025/2026
OECD Employment Outlook 2024
IMF World Economic Outlook — AI and work (Apr 2025)
Banca d’Italia Annual Report 2024
Unioncamere Excelsior 2025
CEDEFOP Skill Anticipation 2025
EU AI Act (Regulation EU 2024/1689)

🎙 Listen to the special podcast (~5 min): https://archive.org/download/ai-breakfast/AI_Breakfast_2026-08-16.mp3

DEV Community