Deepfakes

OpenAI’s Sora 2, an AI video generation tool, is rapidly gaining popularity, sparking debate about deepfakes and content moderation. Despite safeguards, users are finding ways to circumvent them. While lauded for realism and coherence, concerns arise over copyright and potentially disturbing content. OpenAI justifies its approach with a focus on transparency and commercial momentum, amidst growing competition from Meta, Google, and others. The company plans significant spending to expand infrastructure and develop next-generation models, aiming to lead in AI innovation and achieve artificial general intelligence.

In 2024, women in Minneapolis discovered a male acquaintance used the AI platform DeepSwap to create non-consensual, explicit deepfakes using their Facebook photos. This impacted over 80 women and highlighted the growing threat of easily accessible “nudify” apps. The victims are advocating for legislation in Minnesota to curb these services, as existing laws are insufficient. The deepfakes caused significant psychological distress. The case underscores the need for legal and ethical frameworks to address AI misuse while balancing innovation and individual protection. The origins of DeepSwap are unclear due to inconsistent information.