Happy New Year 2026: 2025 Results & 2026 Focus - NAEN/OERN
20.12.2025
133
Greeting • 2025 Results • 2026 Focus
December 2025
Dear colleagues, partners, and friends!
Congratulations on the coming New Year, and thank you for your trust, our joint work, and professional integrity — in complex systems, it is worth more than any loud words.
2025 became a year of acceleration: research outcomes and technological solutions are appearing faster than many organizations can update methodologies, regulations, and quality control. This is not a reason to panic. It is a reason for discipline. Below is a practical review: what got stronger, what got more dangerous, and how to stay the course in 2026.
Shifts in 2025 Worldwide Good / Bad Implications for Subsoil Use Expert Community Summary 2026 Focus
Three words for 2025:
speed verifiability resilience
Speed became the basic norm. Verifiability became the currency of trust. Resilience became what separates the impressive from the working.
Signal #1
Speed became the standard
The «idea → demo → deployment» cycle has noticeably shortened. Slow management turns into a direct risk.
Signal #2
Data quality became an asset
Verifiability, traceability, and change control are a competitive advantage, not a formality.
Signal #3
Security must be built in
Cyber risks, bio risks, and model risks are part of management and accountability, not a separate function of a single department.
Signal #4
Trust is no longer given in advance
Capabilities grew — and so did fakes, errors, and convincingly packaged untruths.
Signal #5
Metrology is back at the center
The more complex the methods, the more important calibrations, protocols, version control, and independent verification become.
Key takeaway of 2025: trust now has to be engineered — just like infrastructure, methodologies, and regulations. If you don’t engineer it, it breaks on its own.
2025 in the world: what shifted and why it concerns everyone
If you look at the key scientific and technological signals of 2025, the picture is simple and uncomfortable: the same tools simultaneously increase productivity and increase risk. Artificial intelligence accelerates research and engineering work — and at the same time makes fakes and attacks cheaper. Biotechnology makes treatment more precise — and demands a stricter culture of biosafety. Fundamental science expands the boundaries of the measurable — and raises the bar for expertise and verification.
AI — from assistant to co-author
signal of the year
AI has become a working tool: it helps find vulnerabilities, speeds up analysis, and supports formalization and verification. But it also symmetrically strengthens offensive capabilities.
- automation in cybersecurity has grown for both defense and offense;
- machine verifiability is gradually becoming the standard where the cost of error is high;
- labor reallocation is accelerating: routine analytical contours are automated first.
Medicine — closer and more precise
practice
The share of technologies that reduce invasiveness, speed up diagnostics, and shift therapy toward personalization is growing. The overall logic has become stricter: first data and protocol, then promises.
- diagnostics and primary patient triage are becoming faster thanks to digital tools;
- new classes of therapy are emerging, including gene approaches and targeted interventions;
- prevention and early detection programs deliver system-level impact.
Materials, energy, measurement
needs control
New materials, ultra-precise measurements, and energy-technology prototypes expand the engineering menu of the future. The entry price is expertise, quality control, and strict verification procedures.
- new classes of materials and states of matter are emerging;
- instrumentation is accelerating and miniaturization continues;
- data volumes are growing, and so is the cost of errors in interpretation.
The honest formula of the year
The faster solutions appear, the more expensive the absence of verification procedures becomes. In 2025, this was especially noticeable in examples ranging from AI and cyber resilience to biomedicine and the instrumentation base.
2025 results: good and bad
Good
Practice signals worth factoring into strategy
1. AI and digital infrastructure
speed • engineering • security
- autonomous systems in pilot modes show lower crash rates within their operating conditions;
- AI in cyber defense speeds up vulnerability discovery, telemetry analysis, and patch preparation;
- formal methods and machine verification strengthen the areas where the cost of error is highest;
- complex intellectual tasks are increasingly solved faster through a human + tool pairing.
2. Medicine and biotechnology
accessibility • precision • quality of life
- digital diagnostics and decision support speed primary patient screening and lower the threshold for accessing care;
- therapy and pain-management options are expanding, and safer approaches are emerging;
- in infectious diseases, the focus is strengthening on effectiveness and on fighting drug resistance;
- gene-therapy and targeted interventions are moving toward reliable clinical track records.
3. Energy, materials, instruments
engineering • metrology • long horizon
- new materials are moving into sensors, coatings, and quality control methods;
- energy prototypes expand options but require mature regulation and strict verification;
- instrumentation is becoming more precise and mobile, so data flow grows and so does responsibility for data quality;
- fundamental observations refine models of the world and discipline thinking.
Bottom line: speed became the norm, but the winners are those who increase speed without losing verifiability and accountability.
Bad
Risks that grew faster than the culture of control
1. Cyber risks and drift in AI contours
attacks • self-change • unpredictability
- AI safety is not fixed once: settings, data, and fine-tuning can change model behavior;
- malware becomes more adaptive, and attack tools get cheaper and faster;
- an organization without an update and access-control process remains vulnerable even if it has modern tools.
2. Trust in data and research
fakes • errors • reproducibility
- generative fakes have become more convincing, including in images and visualizations;
- errors and bad faith undermine the chain «research → decision → investment»;
- reliability is increasingly defined by procedure, not by the authority of a source.
3. Bio risks and ecosystem shocks
dangerous practices • fragile systems
- in biology, the cost of error can be disproportionate to the benefit, so biosafety becomes a management discipline;
- natural systems break nonlinearly: degradation can be massive and fast.
Sober conclusion:
naive belief in the automatic truth of data, models, and nicely formatted reports in 2026 will be costly — in money, time, and reputation.
Expanded summary
Opportunity review: where you can speed up without losing control
AI and infrastructure: the upside
Routine contours are accelerating: error search, log analysis, initial classification, anomaly control, and material preparation. With a properly built process, this saves time and reduces the number of misses.
AI and infrastructure: the risk
The most expensive mistake of 2026 is believing that a tool replaces a process. AI strengthens discipline, but does not substitute it. Without regulations, logging, data control, and accountability, AI will accelerate not only work, but also errors.
Medicine and biotechnology: evidence is getting more expensive
In biomedicine, a beautiful result has long been different from a working solution. Protocols, reproducibility, real-world safety, and time-based follow-up are needed. The same principle carries over to any field with a high cost of error: the quality of the outcome is determined by the quality of the process.
Where data becomes digital and massive, the winner is not the one who promises louder, but the one who proves more carefully.
Energy, materials, and instruments: the fundamental becomes applied quickly
New materials and measurement methods almost always seep into practice through instruments, sensors, quality control, and standards. That is why long-horizon industries, including subsoil use, should stay connected to the scientific agenda — without cults and without illusions.
The main filter is reproducibility. The more complex the method, the more it must be independently verifiable.
Quiet-scale wins
- prevention and early diagnostics deliver the most sustainable effect where infrastructure and quality rules are built;
- lower treatment costs and broader access to essential medicines strengthen system resilience;
- disease control programs show that process-driven wins are possible when discipline is stronger than noise;
- long-term trends improve not because of slogans, but because of management: data, standards, accountability, and verification.
Bottom line on opportunities: think faster, verify faster, deploy faster — and still keep accountability.
Risk review: what really works in 2026
The unpleasant truth about security
Only process works. A product without a process is a display window. The minimum set of measures includes: asset inventory, least-privilege access, logging, updates, segmentation, staff training, and an incident response plan.
AI needs a separate governance layer: usage rules, a ban on use in critical contours without validation, recording data sources, and protocols for independent verification of outputs.
What reliable evidence looks like
- primary data is available at least for audit and protected from silent edits;
- traceability exists: from data to conclusion — step by step, with a change log;
- reproducibility is ensured: an independent party can rerun the calculation;
- assumptions and uncertainty are described honestly: ranges, sensitivity, limitations;
- independent review is built into the process, not added at the end as a formality.
Stop list for 2026
- taking results on faith without primary data and a change log;
- making critical conclusions from pretty pictures without verifying origin and quality;
- using AI without rules: where it is allowed and where it is not, who is accountable, how sources are recorded;
- treating security as one department’s job: resilience is shared accountability;
- hiding uncertainty behind precise numbers: that is not precision, it is self-deception.
Fact: in 2026, the market will increasingly pay for verifiability, not persuasiveness.
A small bonus of 2025: professional curiosity is useful
Practice requires discipline, but motivation feeds on curiosity. The world remains complex — and that is good news: a complex world always has something to learn.
- astronomy and physics add new observations and refine models;
- biology regularly brings data on behavior and communication in living systems;
- instrumentation makes it possible to measure what recently seemed out of reach;
- interdisciplinary approaches speed the transfer of fundamental knowledge into practice.
Practical takeaway: curiosity is the fuel of competence. But competence without verifiability does not withstand reality today.
What this means for subsoil use and expert review
Subsoil use is a long-horizon discipline where mistakes are expensive. In 2025, the standard of what is acceptable changed: competence alone is no longer enough — you must be verifiable. Having data is no longer enough — you must be able to prove its origin, quality, and immutability. Deploying digital tools is no longer enough — you must manage their risks.
Five shifts that will affect everyone
- Audit trail as the norm. From field measurements and samples to the final conclusion, the chain must be reconstructable step by step.
In practice: metadata, version control, processing protocols, and recorded changes.
- Images and interpretations are a risk zone. As fakes increase, expert review must detect artifacts and anomalies.
In practice: standards for core photos, thin sections, and microscopy; origin control; manipulation checks.
- Cyber resilience is part of project resilience. Software and infrastructure vulnerabilities hit production and reputation.
In practice: segmentation, backups, an update regimen, training, and access control.
- AI is a tool, not magic. It speeds analysis, but requires applicability rules, validation, and accountability.
In practice: testing on reference cases, recording sources, banning black-box use in critical conclusions.
- Reproducibility is a competitive advantage. Those who can prove data and calculation quality win the trust of markets and regulators.
In practice: independent recalculation, standardized reporting, and honest uncertainty disclosure.
Mini checklist for 2026 for a project lead
1. Data: unified formats, version control, change log, primary data storage, access rights.
2. Measurements: calibrations, protocols, quality criteria, error descriptions, applicability limits.
3. Calculations: reproducible models, fixed parameters, repeatable scenarios, control of initial assumptions.
4. Expert review: independent review of key conclusions and bottlenecks before final delivery, not after.
5. AI: usage rules (where it is allowed and where it is not); testing on typical cases; recording sources and outputs.
6. Security: least-privilege access, staff training, incident response plan, backups.
A simple maturity test
If an independent expert cannot quickly understand where the data came from, how it was checked, which assumptions were applied, and how to reproduce the calculation, then it is not a document, it is a presentation. In 2026, presentations stop working where the stakes are high.
Where AI is useful already — and where it cannot be allowed without safeguards
Where AI delivers real value
- anomaly detection in large datasets: geophysics, monitoring, telemetry;
- initial classification and labeling: core, thin sections, images, documents;
- routine acceleration: reporting tables, draft descriptions, completeness checks, inconsistency detection;
- a second opinion for experts with transparent sources and clear limitations.
Where AI is dangerous without strict rules
- final reserve and risk conclusions without independent validation and reproducible calculations;
- generating «evidence» without controlling origin, primary data, and audit;
- handling sensitive data without a threat model and access regime;
- replacing expert judgment with model authority — a direct path to expensive mistakes.
Rule for 2026: AI is allowed into critical contours only where there is verification, logging, and accountability.
2025 results for the expert community: discipline, methodology, trust
For the Association of Organizations in Subsoil Use "National Association for Subsoil Expertise" 2025 was about strengthening practices that sustain trust: methodological rigor, independent expert position, reproducibility, and professional ethics. We deliberately moved from declarations to procedures — because procedures outlive technology shifts and trends.
Methodology
Focus on evidence, proper work with uncertainty, and transparency of assumptions. In the real economy, this reduces the risk of expensive illusions.
Standard: «you can see how it was done» matters more than «it looks good».
Qualifications and accountability
A culture of competent persons: a signature and a conclusion are not a formality, but personal accountability for quality and logic.
Standard: accountability for the conclusion outweighs accountability for the form.
Digital transparency
Developing an infrastructure of knowledge and practices so the market understands how it was done, not only what was claimed.
Standard: version control and change traces are part of the professional standard.
Expanded focus points for 2025
- Traceability. Any key conclusion must unfold into a chain: data → processing → interpretations → calculation → conclusion.
- Uncertainty. Do not hide uncertainty behind precise numbers; describe ranges, scenarios, and sensitivity.
- Independent review. Where the stakes are high, independent verification is savings, not expense.
- Data culture. Data quality is a managed system, not a one-off check before delivery.
- Readiness for new tools. AI and automation must increase reproducibility and speed, not add a black box to the critical contour.
Principle: speed is acceptable only where it does not destroy verifiability.
2026 focus: discipline, transparency, technology
In 2026, the winners will be those who do not argue with reality. Reality is this: speed increased, risks increased, and data quality became a key factor of trust. That is why the focus must be practical and measurable.
1. A unified standard of evidence
Clarify requirements for source data, assumptions, uncertainty, and the form of results so that independent verification is fast and honest.
2. Data quality control by default
Verification, version control, logging, independent validation — not optional, but a baseline norm.
3. AI as an expert tool
Use AI where it improves reproducibility and speed, and restrict it where it creates a black box in critical contours.
4. Ethics and security
Zero tolerance for falsification, careful handling of sensitive data, cyber hygiene, and readiness for incidents.
5. Practice over declarations
Fewer promises — more regulations, protocols, training cases, and clear quality criteria.
A practical sequence of actions in 2026
Step 1
Fix the critical contours: which data and conclusions are expensive, where the primary data is, who is accountable, where vulnerabilities and substitution points are.
Step 2
Build verifiability in: versions, change logs, a calculation reproduction template, independent verification of bottlenecks.
Step 3
Automate and apply AI where the risk is minimal and the effect is maximal: anomalies, classification, completeness control, material preparation.
Anchor point for 2026: professional expertise must be both deep and verifiable. Everything else is noise.
I wish you resilience, clarity in decisions, and professional wins in 2026. May your data be clean, your models verifiable, your partnerships honest, and your results worthy of publication and time.
Sincerely,
Andrey Viktorovich Tretyakov
Director, AOON "NAEN", Member of the "OERN" Board.