From the same wonderful people who brought gold farming bots to life, online poker bots may soon come online to harass online poker players according to the Freakonomics blog. The difference between bots of yore and the potential new breed of bots is that rather than just replicate mundane in-game tasks, poker bots will actually have to outplay their opponents. However, once poker is "solved" like checkers was "solved," it would seem the bot should be able to beat humans over the statistical long run, although any given hand may go one way or the other based on the simple "luck of the draw."
As the article points out, poker sites claim they will boot bots. This may prove more complex than anticipated. Randomization can be incorporated into bot play, or a human may simply reference the bot's choice and play accordingly. Penalizing players who do "too well" over the long term isn't a viable option either, as a good player can simply do well on the long term.
There is one point of contention I take with the article, and that is the classification of poker as "pure skill" like "chess and checkers." This is simply not the case. While poker has a defined and significant skill element, there is still a draw pattern that is not any result of player skill. Chess and checkers have no randomization at all. The pieces always start in the same place, and both players start with identical position. While poker does have skill, the simple dealing of cards at the beginning introduces an element of chance. While I don't wish to argue the varying levels of chance vs. skill in this post, I cannot agree with anyone who classifies poker as a "pure" chance or "pure" skill game. It contains elements of both.
Bots haven't toppled other online games, but the poker bot theory may provide the greatest anti-bot challenge to date. Only time will tell how online poker sites decide to deal with bot players, but I doubt that bots will end online poker.