Online: Trending Now

Technologists, economists, and visionaries are warning us that in the next three to 18 months, we are going to experience rapid and pervasive disruption of our professional lives, workplace models, and distribution of income. Professional positions requiring college degrees will be lost, remade into highly-productive, cost-efficient, hybrid human-AI models where human contributions and compensation will collectively shrink and evaporate.

The question we must answer very soon is what can we teach that will prepare our learners to endure the huge changes that are upon us? We must not “stay the course” as it becomes abundantly clear that things are not going to be the same. Lest we say that these forecasts and claims are mere hyperbole, let’s examine the GDPval benchmark report “Measuring the performance of our models on real-world tasks” indicating AI models are approaching or matching industry expert performance in complex tasks like spreadsheet modeling and document editing, with significantly higher speed and lower cost than human professionals in 44 professions.

GDPval is OpenAI’s evaluation benchmark, dated September 3, 2025, for economically valuable, real-world knowledge-work tasks. According to OpenAI, this version spans 44 widely distributed professional occupations drawn from the top 9 industries contributing to the U.S. GDP, with a total of 1,320 specialized tasks in the full set and 220 tasks in the “open-sourced gold set.” OpenAI says these tasks were built and vetted by experienced professionals in those fields and are intended to measure performance on realistic work products rather than narrow academic benchmarks.

GDPval is intended to estimate how well models perform on work that resembles what professionals actually do, such as drafting legal briefs, preparing accounting materials, building schedules, generating diagrams, and similar outputs. OpenAI suggests it as a way to track model capability on economically meaningful tasks, not just standard benchmark questions. This means that such applications of AI are likely to lead in implementation in the next few months because of the savings it promises to those entities that use AI to replace previously salary-compensated humans.

OpenAI also indicates that GDPval is part of a broader effort to understand AI’s labor-market relevance and practical usefulness. The public overview page emphasizes that realistic occupational tasks provide a clearer signal of how models may support day-to-day professional work, while also noting limitations and future iteration.The applications are now upon us – they are not merely vague speculation, rather they are near-term, cost-saving implementations that we are seeing installed today.

I asked Gemini 3 Fast Mode to address the reality so far in 2026:

We are currently in a period of “Job Redesign” rather than total elimination. 91% of enterprises report that while they haven’t eliminated entire departments, the nature of the roles within those departments has fundamentally changed. The primary evidence today isn’t a high unemployment rate (which remains relatively stable at 4.28%), but a widening wage gap between those who can “pilot” AI and those whose tasks have been fully automated.

Yet, there is evidence of which professions as well as people are more likely to persevere rather than be laid off. In a very recent Washington Post article Kevin Schaul and Shira Ovide write:

As artificial intelligence becomes more capable, some jobs may remain in demand while others decline. Web designers and secretaries are more at risk than janitors, according to one recent study. But there’s another dimension to the picture. Some workers will find it easier to adapt, the researchers argue, based on factors like their savings, age and transferrable skills… Women make up about 86 percent of those most vulnerable workers, the researchers said, suggesting the negative effects of automation won’t be borne equally across society.

An extinct occupation, telephone switchboard operators, offers some possible reasons for both hope and pessimism about AI’s effects. It was once one of the most common jobs for American women, but jobs were wiped out as telephones modernized starting in the early 20th century, according to a research paper published in the Quarterly Journal of Economics in 2024 by American professors James Feigenbaum and Daniel Gross. Switchboard operators who lost their jobs were far more likely than their peers to never find other work or to take lower-paying jobs, the research found. But within years, new opportunities opened for young women as secretarial and restaurant work boomed. “I read that as somewhat hopeful,” Feigenbaum, a Boston University economic historian, said in an interview.

To survive the likely disruption in needed workplace skills that is rapidly overtaking our workplaces, Gemini 3 Thinking identified four key durable personal traits:

• • Metacognitive Agility: The ability to “learn how to learn” and the humility to remain a “rookie” as tools evolve every few months.
  • Ethical Discernment: AI can provide options, but it lacks a moral compass. The ability to weigh the societal and human consequences of a decision is a premium trait.
  • Empathetic Leadership: Managing teams in an era of uncertainty requires high emotional intelligence (EQ) to maintain morale and navigate complex human conflicts.
  • Systems Thinking: The capacity to connect dots across disparate disciplines (e.g., law, tech, and sociology) to solve “wicked” problems that AI sees only in fragments.

Gemini 3 Thinking specifically recommends: “Colleges and universities should move beyond basic AI literacy toward Agentic Fluency.” Here is what Gemini 3 recommends we should be teaching:

• • Human-AI Orchestration: Teaching students to move from “prompting” to “orchestrating” autonomous AI agents. This involves managing workflows where the human acts as the creative director and quality controller.
  • Advanced Discernment & Verification: With the rise of synthetic media, we must teach “adversarial thinking”—the ability to question, verify, and contextualize information produced by LLMs.
  • Interdisciplinary Problem-Solving: Removing silos so that a nursing student understands data privacy and a business student understands algorithmic bias.
  • Strategic Communication: While AI can write a memo, it cannot build an authentic relationship. We must emphasize face-to-face negotiation, high-stakes presentation, and persuasive storytelling.

    The goal is to produce ‘AI-proof’ graduates who don’t compete with the machine, but rather provide the human ‘why’ behind the machine’s ‘how.’

Industry leaders are telling us that the technology, infrastructure and demand are beginning to accelerate at exponential scale. The convergence of compute, energy, intelligence, embodiment, and orbital access into one compounding system is coming. Futurist Peter H. Diamandis notes in his Metatrends Substack: ”That future isn’t ten years away. It’s arriving now and deploying over the next 12-24 months.”

We in higher education have a responsibility to our learners to prepare them for the workplace of today and tomorrow. What are you doing to ensure that the degree and certificate candidates at your institution are prepared to thrive in the new environment that is emerging in this year and next?

 

This column was originally published in Inside Higher Ed.