In three years, artificial intelligence has become the most rapidly adopted technology in history and is changing the rules in both industry and services. A lecture by a computer science professor summarized what is worth focusing on in 2026: autonomous agents, fusion models, AI factories, and the rise of intelligent robotics. It also warned about the risk of digital exclusion for people who cannot keep up with the pace of technology.
2026: Autonomous Agents and Fusion Models
Autonomous agents are systems that independently perform tasks without constant human intervention. A practical example is planning a business trip: the AI checks the calendar, chooses a flight and hotel, verifies internal policies, and fills out the travel authorization. Such capabilities save time but raise questions about what junior positions will do as their routine tasks are automated. The market and the education system will have to respond quickly by finding new roles for young professionals.
Fusion models combine two or more specialized models into one so that the result is neither slow nor unnecessarily large. It is a technique where parameters are aligned, overlapping parts are merged, and redundant ones are removed. It is expected that in 2026 the first strong applications and startups will make this practically usable. The speaker’s team has announced a publication on this topic in the first quarter of the year.
Robotics Enters the Physical World: From Lights-Out Factories to Homes
"Lights-out factories" already exist: fully automated operations where robots assemble, package, and move goods without the need for human presence. Such a model can significantly affect carmakers in Slovakia as well and reconfigure job roles across entire regions. Intelligent robotics pushes AI from the screen into reality—the machines orient themselves in space and handle delicate tasks. All this indicates the rapid maturation of technologies that until recently we saw only in videos.
After visiting the two largest Chinese manufacturers, the speaker states that humanoid robots already know how to walk, grasp objects, and not fall, although the hardware and the connections to AI models are still being fine-tuned. Investment in development is massive, so rapid progress can be expected. According to his estimate, the first units will appear in homes by late 2027–2028, and within 5–7 years they may be common among the upper middle class. At the same time, society must prepare for changes in work and the need for new skills.
AI Infrastructure, Sovereignty, and the Quantum Horizon
AI "giga factories" are halls full of computational accelerators intended for training and operating models; the problem is that the hardware needs to be replaced roughly every two years, so the largest cost is ongoing renewal. Nevertheless, it makes sense to build infrastructure at home—because of data sovereignty and application security. In Slovakia, Fakulta informatiky a informačných technológií STU with SAV and partners within EuroHPC is launching projects, for example AI Antenna for transferring European capacity and supporting small and medium-sized firms. There is also a growing AI competence center for education and a new master’s program in English.
Quantum computing is still not very powerful, but it has a clear roadmap: around the turn of the next decade it could handle tasks that are extremely demanding for classical machines today. To break common asymmetric ciphers would require on the order of 100,000 qubits; today’s machines have only thousands, but the direction is set. If this progress is confirmed, quantum AI could handle some heavy computations more efficiently than large AI farms. That would change the cost equation and infrastructure needs in favor of smarter, not just bigger solutions.
