Come on people, apply some basic statisticsto this. Think about it this way, if you have enough accounts at your brokerage, even if every one was controlled by monkeys throwing darts at a list of stocks, then some of the accounts by pure random chance would show outsized returns. And you'd be an absolute moron to copy those account's trades going forward based on that observation. This is effectively the definition of data mining! I've got hundreds of backtest that showed spectacular results over months and even years that failed as soon as I did an out of sample test, this is exactly that only a completely naive version of it.
There may be a dipshit employee at a brokerage who doesn't grasp basic statistics, doesn't understand how the brokerage makes its risk free money from commissions and the idiocy of trading that for trading, and doesn't care about the significant penalties they face. But it makes no sense for the brokerage itself to engage in this anyone who does is enough of a moron you probably don't need to worry about them reverse engineering anything!