Unless war, natural disasters or other major global disruptions knock humanity off course, the coming decade promises an extraordinary period of technological change.
The next decade could bring one of the most consequential periods of technological change in modern history. Not because one invention suddenly changes everything, but because artificial intelligence, robotics, biotechnology, advanced computing and energy systems are beginning to interact. The question is not simply what these technologies can do. It is whether our institutions and our people will be ready for the consequences.
An inflection point, not a timetable
An inflection point occurs when the direction or pace of change shifts markedly. We cannot know whether 2026–2036 will be that period. Yet it is reasonable to prepare for the possibility. Some developments are visible today; others are projections; a few remain speculative. Treating all three as certainties would be as unwise as ignoring them.
AI moves from answering questions to taking action
Today, much AI use still involves producing information for a person to assess. The more profound development may be systems that carry out connected tasks: reviewing documents, coordinating suppliers, writing software, booking services or managing limited expenditures. In some workplaces, the human role could move from performing every stage to specifying outcomes and supervising exceptions.
My prediction: by the early 2030s, organisations will routinely use bounded AI agents for substantial business processes. The main obstacle may be less the intelligence of the agent than the reliability, permissions, audit trails and clear assignment of responsibility needed to trust it.
Robotics brings AI into the physical world
Robots are already established in industrial production, and service robots are spreading in commercial settings. Advances in perception, language interfaces and manipulation could allow machines to operate in more variable environments. Warehouses, hospitals, agriculture, cleaning, inspection, construction and delivery are possible growth areas.
My prediction: specialised robots will spread faster than general-purpose humanoid household servants. Humanoid systems may gain a foothold in structured workplaces, but cost, safety, maintenance and dexterity will limit universal deployment. A machine that performs a narrow job reliably may be commercially more important than a robot that looks like a person.
Quantum computing: powerful, specialised and probably remote
Quantum computing is not simply a faster version of the desktop computer. Its potential lies in specialised classes of calculation, including selected problems in chemistry and materials science. Much leading hardware requires specialised cooling and control; access through remote computing centres is more plausible for most users than a quantum computer on the desk.
IBM currently aims to deliver its Starling fault-tolerant system in 2029, but roadmaps are targets, not guarantees. My prediction: useful quantum services may grow during the 2030s, initially for restricted problems. Claims that quantum computing will transform every business task by 2030 should be treated cautiously.
Medicine and biotechnology
AI is already helping researchers identify candidate medicines and analyse biological data. Combined with automation in laboratories, improved imaging and genetics, it could accelerate portions of research and diagnostics. But clinical trials, safety, biological complexity and regulation will remain essential.
My prediction: the decade will bring meaningful improvements in selected diagnostics, drug discovery and personalised care, though not a sudden cure for most diseases. A particularly consequential development may be the integration of AI, laboratory robotics and continuously updated evidence.
Energy, infrastructure and materials
More computing requires more electricity, data centres, grids, cooling and physical resources. The International Energy Agency projects substantial growth in data-centre demand. This creates opportunity for grid modernisation, storage, better power management and new generating capacity, but also potential delays, costs and environmental pressures.
My prediction: energy availability and grid connections will become a practical constraint on some AI expansion. Progress in batteries, power electronics, materials and renewable generation may help, but there is no reason to assume that a speculative energy breakthrough will solve every bottleneck.
Transport, security and defence
Autonomous systems may expand first where conditions can be controlled or well mapped: ports, mines, depots, logistics routes and industrial premises. Public roads and densely populated settings will remain harder. In parallel, AI will make some cyberattacks, misinformation campaigns and surveillance more scalable, while also strengthening detection and defence.
My prediction: the struggle to secure autonomous and AI-connected systems will be a defining governance problem. The more authority a system has to act, the more important are independent checking, secure permissions, records of decisions and the ability to stop it.
Work and education: the less visible transformation
A technology need not eliminate an entire profession to disrupt employment. It can remove the junior tasks through which people traditionally learned to become experts. It can also change the demand for administrative, technical and creative work. Some new jobs will appear, but the transition may be uneven, and retraining will not automatically replace lost opportunities.
Education faces a double requirement: teach pupils and students to use AI effectively, and preserve their ability to research, reason, create and judge without it. Assessment should establish what the learner contributed and can demonstrate independently. A useful course may deliberately combine AI-assisted and non-AI work.
The reinforcing cycle — and its limits
The most speculative possibility is an accelerating feedback loop: AI helps design chips, robots and materials; robots and automated laboratories produce better data; improved computing supports stronger AI. These systems could reinforce one another and shorten development cycles. But physical experiments, engineering constraints, finances, regulation and societal acceptance will continue to impose friction. Exponential progress cannot simply be assumed.
The central question: who remains capable of command?
As systems become more dependable, people may intervene less often. That sounds like progress, but it creates a paradox: the human may be expected to take command precisely when they have had the least recent practice. Aviation and other highly automated domains show why training, competence and meaningful intervention matter.
Technology policy therefore needs more than rules written at the start of deployment. Organisations require operational assurance: specifications, independent checks where justified, monitoring at runtime, clear limits on autonomous execution and human authority to pause, redirect or stop the system. Above all, they need continuing human capability.
Preparing for 2036 without pretending to know 2036
Governments, schools and organisations should plan against several plausible futures rather than one dramatic prediction. They should test systems on bounded tasks, invest in infrastructure and cybersecurity, examine workforce pathways, preserve independent skills and keep responsibility visible. The greatest error may be to confuse possession of a powerful tool with readiness to use it wisely.
The coming decade may or may not prove to be the inflection point. But preparing for the convergence of these technologies is already a rational decision. The central question is not whether machines will become more capable. It is whether people and institutions will remain capable of directing them.
Sources
- IBM Quantum roadmap (updated 2026). https://www.ibm.com/roadmaps/quantum/2030/
- International Federation of Robotics (IFR), 30 September 2026, “Global Sales of Professional Service Robots Surge 24%.” https://ifr.org/ifr-press-releases/global-sales-of-professional-service-robots-surge-24-percent
- International Energy Agency, Energy and AI (2025). https://www.iea.org/reports/energy-and-ai
- Nature Medicine (2025), AI-enabled drug discovery clinical milestone. https://www.nature.com/articles/s41591-025-03832-2
- Nature Reviews Drug Discovery (2026), AI in drug discovery. https://www.nature.com/articles/s41573-026-01496-2
Further articles and books by R T Scanlon: ai-humancapability.com