HUMAN PLUS MACHINE

Recent results from entrance exams at Tokyo and Kyoto universities, where the ChatGPT 5.2 Thinking model not only passed the threshold but outperformed the best applicants, served as a kind of wake-up call for humanity. Just two years ago, in 2024, neural networks were unable to meet even the most basic requirements of Japanese universities. However, today they demonstrate knowledge levels comparable to the academic elite.

What is next?

It is evident that this event will soon render the traditional education model obsolete, which placed a high value on the accumulation of data in the human mind. If information can be structured in seconds, does it still make sense to be a 'walking encyclopedia'?

However, an analysis of the data indicates that the machine has not, in the broadest sense, become truly 'smarter'. Instead, it has achieved impressive efficiency in data processing and logical operations. The fact that artificial intelligence (AI) scored top marks in mathematics and excelled in natural sciences demonstrates that if knowledge can be structured, digitised, and transformed into a chain of logical deductions, the machine can do this better, faster, and without the errors caused by human fatigue.

However, it is important to address the areas where AI has not met expectations. In the humanities. In this instance, the system demonstrated significantly lower performance in tasks requiring extended answers. This underscores a significant gap: current algorithms are yet to fully grasp the 'cultural code', which remains accessible only to humans, according to AI developers. While machines can process factual information about world history, they are not equipped with the capacity for historical empathy and understanding of cause-and-effect relationships that extend beyond statistical probabilities.

A pertinent question to consider is the impact of artificial intelligence on the professional landscape, particularly in terms of the future of human employment. Current developments in machine learning and AI have reached a point where these systems can not only replicate, but also outperform, human intellectual capabilities in various academic disciplines. This raises significant questions about the future of employment, particularly in academic settings where AI-generated outputs often surpass those of human students. This scenario poses a challenge to the traditional educational paradigm, where the value of education is often determined by the skills and knowledge acquired through hands-on learning and practical experience.

Risks for professions

Reports from the World Economic Forum indicate a shift towards a 'judgement economy', where the primary value lies in the ability to operate effectively in uncertain environments, as opposed to the accumulation of knowledge. In the field of medicine, for instance, the role of the doctor evolves from that of a mere 'repository of manuals' to that of an 'architect of treatment'. Algorithms are used to analyse and decode data, allowing doctors to focus on patient care, understanding the psychosomatic context, and making well-informed decisions that are within their legal and moral responsibility. In the field of architecture, AI will optimise drawings and facilitate the creation of large-scale projects. However, it is the human ability to define the 'spirit of the place' that remains unparalleled. It is evident that junior employees in law and related professions will be at risk of losing their jobs to archive automation before they have had the opportunity to gain the necessary experience. Analytical work such as contract review, precedent searching, and risk assessment is already being delegated to neural networks. Experts have expressed concerns about a potential 'competency gap'.

If education does not adapt to include teaching strategic thinking, managing complex social interactions, and creatively synthesising data from various fields, there will be a surplus of specialists who will not be able to compete with algorithms.

System manager

Scientists assert that humanity is entering a 'human-in-the-loop' concept: machines suggest options, humans set goals and bear final responsibility. Consequently, the professional of the future is a meaning operator working with uncertainty. We evolve by freeing ourselves from mechanical labour towards creativity and choice. In accordance with the unanimous consensus of global experts, the replacement of human professionals with machines is not recommended. Instead, the enhancement of human professionals' abilities by training them to become 'system managers' is advised.

This prompts a pivotal question: how should educational criteria in schools and universities be revised, considering the dismal failure of conventional, test-based education to meet the demands of the modern age? In what manner should the assessment of knowledge be conducted? Scientists believe that the answer lies in transitioning to a 'continuous' assessment model, where the focus shifts from memory to cognitive ability. In this context, academia suggests first legalising AI use during exams, turning them into a kind of 'open-code exam'. The process would involve assigning a task to be solved using a neural network. However, the primary objective is not the final answer, but the student's critical report, which would entail identifying logical errors in the machine's work, verifying facts, and proposing improved versions. Assessment focuses not on outcomes but on the quality of human oversight over machines.

It is also necessary to move away from isolated tests across different subjects towards integrated interdisciplinary simulations. Students are tasked with solving real-life problems, for example, a major project that requires simultaneous knowledge of physics, economics, and law. The value of this approach is that it tests not the ability to memorise formulas, but the skill to connect different fields to achieve goals under real constraints.

Assessment should also become dynamic and adaptive. Imagine an examination that adapts its level of difficulty during the course of the process. If a student responds with confidence, the system immediately presents them with more challenging, unconventional questions to assess their mental agility rather than relying on factual knowledge. Concurrently, 'snapshots' of knowledge should be superseded by long-term portfolios amassing work throughout a school term or university semester. This allows educators to observe real thinking dynamics, progress, and how individuals learn from their mistakes over time.

Finally, an invaluable tool remains the viva voce—live discussion resembling thesis defence. This format ensures that cheating or delegating thinking to algorithms is not possible. During such conversations, examiners instantly discern whether students understand core concepts or have merely memorised text. The ability to articulate thoughts immediately is crucial, and live discussion is the most effective method of verifying whether acquired information has been fully assimilated.

In other words, education must transition from the issuance of one-off diplomas to the establishment of lifelong competency portfolios. This is a continuous endeavour to preserve humanity in a world increasingly dominated by machines.

A project for the future

It is crucial to recognise that these assessment methods signify a developmental trajectory — a future-oriented vision. The current system, which incorporates the State Examination Centre (SEC) of Azerbaijan, relies on standardised tests as an objective necessity. This is currently the only method of ensuring transparency and equal opportunity when admitting tens of thousands of applicants en masse. It is not feasible to replace all tests with portfolios on a nationwide basis at this time; this would require significant resources. Such creative approaches are currently limited to elite foreign universities, where the depth of future specialists' thinking must be assessed after initial selection.

In this context, the SEC's decision to introduce a new foreign language exam model by 2027 appears to be well-timed. As SEC Chair Maleyka Abbaszade noted, this approach marks a shift towards 'functional and communicative assessments' aligned with the Common European Framework of Reference for Languages. In summary, the initiative directly responds to the challenges of today. Traditional grammar tests are susceptible to automation; the focus on systematic development of written skills and real communication, which SEC aims to promote, is precisely the kind of task that cannot be delegated to algorithms. This is an ideal example of how state systems begin aligning with global demands of the 'judgement economy'. Rather than focusing on reference data, the emphasis is now on the ability to convey ideas, understand context and engage in live dialogue. However, this transitional period inevitably causes stress as we ask schoolchildren to meet standards with which they are not yet familiar, as the prevailing pedagogy has not yet had the opportunity to prepare them for these expectations. However, this path is essential: The proposed changes to the SEC should signal a change in teaching culture.

Testing first, then creativity

Scientists foresee a hybrid future learning model: first a computer filter for basic knowledge, then a creative or project phase. It is evident that systems aim to maintain stability by incorporating both established practices and new competencies. However, if educational institutions were to adopt communicative methods, incorporating discussions and problem-solving into their daily routine, then 'new' exam formats would no longer be intimidating and would instead serve to demonstrate everyday skills. Stress levels are reduced when learning becomes 'exams in action'.

It is also critical to strike the right balance: schools of the future should not abandon knowledge testing entirely, but must stop making rote learning the sole success criterion. Basic literacy and factual knowledge remain essential foundations. In hybrid models, the computer filter serves as an 'entry ticket' confirming mastery of fundamentals. From a business perspective, the focus should be shifted towards project activities and interdisciplinary synthesis in the learning process.

Teachers transition from mere transmitters of data to facilitators, imparting 'digital hygiene' skills such as working with AI, recognising its 'hallucinations', and personalising it with experience. In such systems, the focus is on students who can apply knowledge collaboratively, show empathy, and find unconventional solutions to real problems, rather than those who can memorise paragraphs perfectly. This approach lays the foundation for future careers.

Azerbaijan has learned a crucial lesson: it must avoid merely copying outdated standards; instead, it must make a technological leap. Investment should be focused on developing critical synthesis skills, interdisciplinarity, and ethics. If Azerbaijani students continue to prepare as they have for decades, graduates will enter professions destined to disappear before they receive diplomas. Azerbaijan has a chance to prioritise AI system managers, ‘architects of meaning’, and new types of doctors. The key conclusion: machines take over routine tasks; humans take responsibility. It is precisely this responsibility combined with factual knowledge that makes professionals whose value will only grow.