Dropbox states AI and autonomous AI agents are used by threat actors to enhance attacks.
“Emerging and evolving cybersecurity threats pose significant challenges and involve increasingly sophisticated threat actors, including those leveraging AI and machine learning. Malware, ransomware, viruses, social engineering (such as phishing), denial of service, insider misuse, and other network attacks have become more prevalent, particularly against cloud-based services. Threat actors may use AI or autonomous AI agents to enhance and scale attacks, including generating convincing phishing content, deepfakes, and adaptive malware, making detection and mitigation more difficult. In this fast-changing threat environment, we are continuously assessing our security posture, including through the use of penetration testing and red team exercises, to identify gaps, threats, and vulnerabilities and, where we believe appropriate, we actively take additional and ongoing steps that are intended to strengthen our defenses, including cybersecurity capabilities, and to mitigate the risk of a breach or incident. We also evaluate risks associated with our use of AI, including risks related to data integrity, model security, and potential misuse or manipulation. However, AI technologies may introduce new or amplify existing risks, including unintended system behaviors or exploitation by malicious actors. If we fail to respond appropriately to any identified gaps, threats or vulnerabilities, including by providing sufficient investment and prioritizing strategic initiatives, or if we fail”