#预测市场 I see this kind of argument again—predictive markets are going to be manipulated by AI, and public opinion becomes hard to discern. Honestly, this concern strikes me as both fresh and very old.

Looking back at history, it’s clear that as early as 1905, The Washington Post warned about manipulation in betting markets; in 1916, the Democratic Party claimed there were "issues" with the election market; in the 2004 Berlin state elections, political parties openly organized supporters to pour money in to boost prices. These stories all tell us a simple truth: as long as there are prices, someone will try to distort them. But this is not the original sin of predictive markets; it’s the market’s own destiny.

What’s more interesting is that history also proves another point—manipulation isn’t that easy. The strange surge of Romney on InTrade in 2012, which fell back within days; the large order on Polymarket about Trump in 2024, later revealed to be a French investor doing private polling to make money. Manipulators repeatedly invest huge sums, but the results are always short-lived waves; market resilience is far stronger than we imagine.

What truly makes me alert isn’t what manipulation can do, but what it can trigger. The moment CNN started broadcasting daily predictions of market prices, the game changed— even if manipulation fails, accusations of "foreign interference" and "elite collusion" can still spread everywhere, and the truth gets drowned in panic. That’s the most dangerous part.

So the solution is clear: media must have bottom lines and only report on markets with sufficient liquidity; platforms need to install monitoring systems; policymakers should take action against obvious manipulation. In other words, it’s not about abandoning predictive markets as a tool—in an era where AI is everywhere and polling response rates are plummeting, we need it more than ever—but about ensuring they operate within an open and transparent framework. History shows us that markets will self-correct; the key is not to let noise drown out signals.