What you are referring to is the Efficient Market Hypothesis (EMH).
It's good to understand that there are generally three forms of this idea:
- Strong: All private and public information is reflected in current stock prices
- Semi-strong: All public information is reflected in current stock prices
- Weak: All historical price information is reflected in current stock prices
Without delving into the technical details of these three ideas, the idea is that markets are effective at price discovery and that the idea that you can find a stock that is mispriced and gain from that 'arbitrage' is impossible or at least infeasible.
You didn't ask if the theory is correct so the direct answer to your question of whether this idea assumes or requires that most participants of the stock market are well-informed is: not exactly.
For one, the weak and semi-strong forms of the hypothesis allow for private information e.g.: insider information to not be reflected in the price. I think we can safely assume that most people don't believe insider information is reflected in stock prices. There are laws against insider trading precisely for this reason. The insider selling or buying something that they know is worth less or more than the majority of market players.
Even beyond that, all of these really only require that there enough informed players to move the price to its correct level. It's also important to realize that all market players do not have the same ability to move a stock price. It's not a one person: one vote system. The average investor cannot move a stock price as much Warren Buffet can, for example.
This has always been my problem with the strong and semi-strong forms: the way that the price moves people are buying and selling based on a discrepancy between the price and the real value. But if the price were already correct, how could that happen? The obvious answer is 'new information' but again, once that new information comes out, there's some period of time when the related stocks have their old (wrong) price. In other words, the way markets discover prices is: when there's a mispriced stock, players will take advantage of that until that discrepancy is resolved which is exactly what the strong and semi-strong theories say can't happen. If the price were always correct, there would be no need for price discovery which is how prices become correct.
In short, other than maybe a few out-of-touch academics, no one really believes this literally true, but I think many people believe it is roughly true over time, as I do.
You may be asking this because there's an idea in economics that humans are purely rational actors with regard to markets and other financial matters. The problem with this idea is that it's known to be empirically false based on many repeatable experiments. One of the main reasons this assumption was used is that it makes the mathematics so much easier. We've only begun to have the kind of computing power required to deal with economic theories that reflect real human behaviors.