The attacks started around Dec. 5 and lasted a few days, sending victims who clicked on the ads to malicious Web pages. Those pages took advantage of known software bugs to install backdoor programs that gave the attackers control of the victims' PCs, or to install software that made it appear as though the PCs were filled with malicious software.

Google acknowleged Friday that it had experienced some issues on its DoubleClick network but said it had put a stop to them quickly.

"[T]he DoubleClick Ad Exchange, which has automatic malware filters, independently detected several [ads] containing malware, and blocked them instantly -- within seconds," Google spokesman Jay Nancarrow said via email. "Our security team is in touch with Armorize to help investigate and help remove any affected creatives from any other ad platforms."

Nancarrow wouldn't say how the malicious ads got onto Google's ad network, but Armorize Chief Technology Officer Wayne Huang [cq] said cybercriminals may have tricked Google by serving the ads from a domain similar to that used by a legitimate ad-serving company, AdShuffle, based in Irving, Texas. AdShuffle couldn't be reached for comment Friday.