We are slowly coming to the end of 2025, another year in which artificial intelligence has played first fiddle. We’re not just talking about the digital world. It is infiltrating literally everywhere, including doctors’ offices, laboratories, courtrooms, grocery stores and even art galleries. Over the past few months, we have seen, among other things, advances in the quality of images and video generated, as well as smarter and more responsible text models. This we already know, but what’s next? Experts at Microsoft have outlined 7 trends in AI for 2026. If we want to keep our finger on the pulse, they are worth reading.
1. Artificial Intelligence will enhance what people can achieve together
Microsoft sees 2026 as an era of new alliances between technology and people. Just as AI has been answering questions and thinking through problems in recent years, its next wave is expected to be based on true collaboration. The future is not about replacing people. It’s about empowering them – argues Aparna Chennapragada, CPO of AI experience at Microsoft.
What does it consist of? The importance of AI agents as digital collaborators who will help small teams measure higher. Chennapragada sees it this way: small teams of up to 3 people, with the help of agents, will be able to launch a global campaign in a few days, with artificial intelligence handling data processing, content generation and personalization. Humans will be able to handle strategy and creativity. It is possible that 2026 will be the year to bring out the best of both worlds.
Aparna Chennapragada has advice for entrepreneurs: Don’t compete with AI, but focus on learning how to work alongside it. The next year belongs to those who elevate humans, not eliminate them. Let’s hope this is not just wishful thinking. After all, we’ve heard about a lot of layoffs in the big tech sector in the last year , and more are planned. All thanks to the fact that AI is cheaper than human workers.
2. AI agents as a workforce will receive new protections
Agent-based artificial intelligence will play an even greater role in day-to-day work, acting more as team members than tools, according to Vasu Jakkal, CVP at Microsoft Security. While organizations will increasingly rely on software to perform tasks and make decisions, the issue of trust in agents will be key. Security is therefore needed.
Every agent should have similar safeguards as humans to ensure that agents do not turn into “double agents” exposing uncontrollable risks, a Microsoft expert argues. This means that each AI agent will be given a clear identity, the scope of the systems and information it accesses will be limited, administrators will be able to manage the data it creates, and that it will be protected from attackers and cybersecurity threats.
Security will be an environment in which agents will be embedded from the start, rather than something added later. In addition to this, the use of AI by cybercriminals is expected to continue to increase, including in new ways. It is therefore important for defenders to be able to use security agents (e.g. Security Copilot in Microsoft Entra) to track threats and respond instantly. Trust is the currency of innovation,” adds Jakkal.
3. AI can narrow the global health care gap
Artificial intelligence in medicine will be at a turning point, predicts Dr. Dominic King, VP of Health at Microsoft AI. As early as 2025, we were showcasing the achievements of medical of Microsoft’s artificial intelligence, which makes diagnoses faster, cheaper and more accurate than experienced doctors. The coming year is expected to bring further flourishing of this technology.
We will see evidence of AI moving beyond the diagnostic experience and expanding into areas such as triage and treatment planning. Importantly, progress will begin to move from research environments to the real world, and new generative AI products and services will be available to millions of consumers and patients.
– Dominic King, MD PhD, VP Health at Microsoft AI.
This progression is also significant in the context of the health care access crisis. According to WHO projections, there will be a shortage of 11 million health workers by 2030. This shortage will deprive 4.5 billion people of access to basic health services. Some of the duties of doctors will be able to be taken over by artificial intelligence – not only in clinical settings, but also at home.
Already, the aforementioned Microsoft AI Diagnostic Orchestrator (MAI-DxO) technology is solving complex medical cases with a success rate of 85.5%, 20% higher than experienced doctors. Microsoft Copilot on Bing, meanwhile, answers more than 50 million health-related questions every day. Dr. Dominic King sees this development as an opportunity to give people more influence and control over their health and well-being. Learn more from our articles:
- OpenAI models will better support mental health and contact with the outside world
- AI assistant in the hospital. Microsoft Dragon Copilot is already helping doctors
- ChatGPT will respond professionally when a user has an emotional crisis
4. AI will stand at the center of scientific discovery
Artificial intelligence is already accelerating discoveries and breakthroughs in domains such as drug design, climate modeling, molecular dynamics or materials design. What’s next? As Peter Lee, President of Microsoft Research, predicts, in 2026 artificial intelligence will not only summarize scientific publications, answer questions and write reports. It will actively participate in the discovery process in physics, chemistry and biology.
AI will generate hypotheses, use tools and applications that control scientific experiments, and collaborate with both human and AI co-researchers, Lee says. AI will become a true member of research and laboratory teams, and will soon, for example, suggest new experiments and even partially carry them out. In this way, the human + AI tandem will come to new discoveries faster, and advances in science will grow even more exponentially.
Sam Altman, CEO of OpenAI, shares a similar view. We wrote about this in July in the article “Sam Altman: AI agents will soon help discover new knowledge”.
5 AI infrastructure will become smarter and more efficient
As recently as 2025, artificial intelligence researchers were highlighting the problem of the ever-growing “AI bubble” that is about to hit the wall. Adding more GPU floors in billion-dollar data centers is not going to guarantee perpetual progress, and eventually another direction will have to be taken – especially if tech giants are thinking about achieving AGI and not just short-term investor satisfaction. Microsoft seems to be aware of this. AI development is no longer just about building more and bigger data centers. The next wave is all about using every ounce of computing power – argues Mark Russinovich, CTO, DCISO and Technical Fellow at Microsoft Azure.
The expert explains that the most efficient AI infrastructure will concentrate computing power more densely in distributed networks. Next year, we will see the emergence of flexible, global AI systems – a new generation of interconnected AI superfactories – that will cut costs and increase efficiency. It is worth mentioning that Poland also has plans to build so-called AI gigafactories. Up to a dozen such facilities may be built in the coming years, and a large part of the funds will be provided by the European Union.
Moreover, the quality of AI is to be measured by the quality of the information it produces, not just its size. This is an interesting addition to Satya Nadella’s earlier statement that he is more convinced by the economic growth that AI will inevitably lead to than by benchmarks. Learn more from our articles:
- Microsoft builds world’s most powerful AI data center
- In 2 years, AI has become 90% cheaper and 2.7 times smarter
- Satya Nadella: the measure of AI success will be economic growth, not benchmarks
6. AI learns the coding language and its context
Software development is literally exploding, with activity on GitHub reaching record levels in 2025. Each month, developers merged 43 million pull requests – 23% more than the year before. In turn, the annual number of commits increased year-on-year by 25%, reaching 1 billion. Also in the past year, AI has moved into an increasingly central place in the software development and improvement process. And what’s next?
As Mario Rodriguez, CPO at GitHub, predicts, a new trend will become apparent in 2026 – repository intelligence. Simply put, it’s about artificial intelligence that understands not only the code, but also the relationships and history behind it.
By analyzing patterns in code repositories, AI will detect what has changed and why, and why the pieces fit together. This context helps it give smarter suggestions, catch errors earlier, and even automate routine fixes. The result will be higher-quality software that helps developers move forward faster – Rodriguez explains.
7 The next stage is closer than we think. AI will join forces with quantum computers
Quantum computers are no longer science-fiction. New solutions are being developed to improve the operation of these amazing machines, and make them more stable, and systems with several thousand qubits are no longer rare. However, it is not quanta alone that are alive. And artificial intelligence is breaking into this disruptive category.
While a world-shaking breakthrough in this field may not be expected in 2026, according to Jason Zander, EVP of Microsoft Discovery and Quantum, we are years, not decades, away from when quantum machines will start solving problems that classical computers can’t handle.
One of the growing trends today is hybrid computing, in which quantum systems work together with AI and “silicon” supercomputers. Artificial intelligence finds patterns in the data, classical supercomputers run powerful simulations, and quantum computers add a new layer that, for example, provides much greater accuracy in modeling particles and materials.
This progress runs parallel to advances in logic qubits, which are physical quantum bits grouped together so that they can detect and correct errors and compute – a critical step toward reliability, the expert adds.
One of the milestones on the road to reliable quantum systems is Majorana 1, Microsoft’s own quantum chip unveiled in 2025. It is the world’s first quantum processing unit (QPU) powered by a topological core, designed to achieve one million qubits on a single chip, Microsoft explains. It is also the only quantum solution designed to catch and correct errors. This architecture paves the way for machines with millions of qubits on a single chip, providing the computing power needed to solve complex scientific and industrial problems.
What’s next? As Jason Zander points out, Quantum advantage will drive breakthroughs in materials, medicine and beyond. The future of AI and science will not only be faster, but will be fundamentally redefined.
Source: Microsoft, own elaboration
