Researchers create ondetecteerbare malware with artificial intelligence
Security researchers have been using artificial intelligence a new type of malware, which does not by antivirus programs is detected.
The hardware was on hackersconferentie Defcon demonstrated by a beveiligingsonderzoeker of Endgame.
The malware searches for legitimate code in software and make small changes, so that an attack can be carried out. If the anti-virus software on a computer if malware can see, there are other changes made. So the malware makes in the long term, a type of aanvalssoftware that is not recognized.
The beveiligingsonderzoeker used to attack an artificial intelligence of the non-protiforganisatie OpenAI, where, among others, Tesla CEO Elon Musk is a part of.
The artificial intelligence had learned with a test in 15 hours how to malware for a computer an unspecified antivirus program could escape. At 60 percent of the attempts was malware not the antivirus software recognized.
Endgame is going to be attack available on ontwikkelaarswebsite Github. By the system to analyze, can security companies to learn how their anti-virus software this could help.