Generative Language Models and Automated Influence Operations: Emerging Threats and Potential Mitigations