Creators of fake accounts and news pages on Facebook are learning from their past mistakes and making themselves harder to track and identify, posing new challenges in preventing the platform from being used for political misinformation, cyber security experts say. This was apparent as Facebook tried to determine who created pages it said were aimed at sowing dissension among U.S. voters ahead of congressional elections in November. The company said that  it had removed 32 fake pages and accounts from Facebook and Instagram involved in what it called “coordinated inauthentic behavior.”

While the United States improves its efforts to monitor and root out such intrusions, the intruders keep getting better at it, said cyber security experts interviewed over the past two days.Ben Nimmo, a senior fellow at the Washington-based Digital Forensic Research Lab, said he had noticed the latest pages used less original language, rather cribbing from copy already on the internet.

Facebook’s prior announcement on the topic of fake accounts, in April, directly connected a Russian group known as the Internet Research Agency to a myriad of posts, events and propaganda that were placed on Facebook leading up to the 2016 U.S. presidential election.This time, Facebook did not identify the source of the misinformation.