According to David Wood (see here), atheists who appeal to the argument from evil are logically inconsistent. Why? Wood offers the following explanation:
For instance, atheists seem to be arguing (1) that human beings are so good that God shouldn’t allow us to suffer, and (2) that human beings are so bad that God shouldn’t have created us (or given us free will, etc.). That is, atheists are simply shocked that a good God would allow human beings to experience all sorts of pain and injustice. “Why doesn’t God intervene?” “Why doesn’t God come down here and protect us?” The point of this criticism is that God should save us from harm (i.e. God is morally obligated to protect us). Therefore, we are worth saving from harm.
(1′) Human beings are so good that God should have created us and shouldn’t have allowed us to suffer.
While (1′) contradicts (2), a major problem remains. The problem is that proponents of the argument from evil need not affirm (1′) or (2). Consider, for example, Paul Draper’s evidential argument from evil based on the biological role of pain and pleasure. Rather than copy the entire argument here, I simply invite the reader to read my summary of Draper’s argument. Notice that neither (1′) nor (2) above follow from any of the premises in Draper’s evidential argument from evil. This is also true of the arguments from evil formulated by Michael Tooley, Bruce Russell, William Rowe, Michael Martin, and Quentin Smith.
Another worry I have about (1) is that it seems to imply that God, if He exists, has moral obligations to human beings only if they achieve some minimum level of goodness. But this seems false, even on the assumption that God, if He exists, has any moral obligations at all. On the assumption that there are any moral obligations at all, a sentient being can be the beneficiary of a moral obligation, regardless of whether that being is generally moral, generally immoral, or somewhere in between.