Experts split on how fast AI will improve in 2026

In short: Predictions for 2026 point to smarter AI and more useful “AI agents,” but there is no clear agreement on how fast this will happen or what it will do to jobs.

What's going on

Some researchers and financial analysts expect a noticeable jump in AI abilities in the first half of 2026. Morgan Stanley points to bigger and more powerful computer systems at leading US labs as a key reason. It also cites claims that “model intelligence” could roughly double, and says newer systems are scoring closer to human experts on certain tests.

A big theme is the rise of “agentic” AI, meaning AI that can take actions over time instead of only answering one question. Think of it like a digital assistant that can plan a trip, keep track of what it already tried, and finish the task later (more like a personal assistant, less like a search box). Analysts also expect fewer mistakes as systems add self-checking steps, and more tools that let people control software by typing plain English instead of writing code.

Businesses are already using AI widely, according to a large NVIDIA survey covering thousands of responses across industries like finance, healthcare, and manufacturing. Another shift is the growth of open-source AI, which means models that are shared publicly so more companies can adapt them for specific jobs.

What to watch

The biggest unknown is pace. Some forecasts expect fast progress driven by more computing power, while others see adoption moving in a normal, gradual pattern and note that US unemployment is still around 4.3%. In 2026, watch whether AI agents become reliable enough for everyday workplace use, and whether job cuts show up broadly in economic data or stay limited to certain roles.

Source: NYTimes