OpenAI disrupted more than 20 foreign influence operations over the past year, according to its quarterly threat report. But there’s no “evidence of this leaning to meaningful breakthroughs in their ability to create substantially new malware or build viral audiences,” the report says.