Chris Lehane, OpenAI's VP of global policy and a veteran crisis manager, faces the challenging task of reconciling the company's stated mission to democratize artificial intelligence with its increasingly perceived actions as another traditional tech giant. Despite Lehane's adept public relations skills, including acknowledging uncertainty and expressing genuine concern for humanity, OpenAI's corporate behavior frequently contradicts its idealistic image.

A central point of contention revolves around OpenAI's video generation tool, Sora, which launched with seemingly copyrighted material. Initially, OpenAI allowed rights holders to opt out of having their work used for training Sora, a practice that deviates from standard copyright norms. Following observed public interest in using copyrighted images, the company shifted towards an opt-in model, leading critics to suggest this was a test of legal boundaries rather than true iteration. This approach has drawn legal challenges from major news organizations like the New York Times and Toronto Star, who accuse OpenAI of using their content for training without sharing economic benefits. When pressed on the issue, Lehane invoked the 'fair use' doctrine, calling it a 'secret weapon' for U.S. tech dominance, while acknowledging the disruptive impact on content creators.

OpenAI's infrastructure demands also raise significant concerns. The company operates data centers requiring massive amounts of water and electricity, with new facilities being established in economically challenged areas like Abilene, Texas, and Lordstown, Ohio. Lehane draws parallels between AI adoption and the advent of electricity, suggesting it will modernize energy systems, but questions remain about whether these communities will truly benefit or merely bear the environmental and financial costs. Video generation, in particular, is noted as the most energy-intensive AI application.

The human cost of AI development is another critical issue. Zelda Williams, daughter of the late Robin Williams, publicly appealed for people to stop creating AI-generated videos of her father, describing such creations as 'disgusting, over-processed hotdogs out of the lives of human beings.' Lehane addressed this by emphasizing processes, responsible design, and government partnerships, admitting that 'there is no playbook for this stuff.'

Further complicating OpenAI's image are its aggressive legal tactics. Nathan Calvin, an AI policy lawyer, revealed that OpenAI sent a sheriff's deputy to his home to serve a subpoena, seeking his private communications related to California AI safety legislation (SB 53). Calvin alleges this was an intimidation tactic, weaponizing OpenAI's legal battle with Elon Musk as a pretext to target critics. He publicly branded Lehane the 'master of the political dark arts,' an accusation that stands in stark contrast to OpenAI's mission of benefiting humanity.

Perhaps most indicative of the company's internal struggles is the growing dissent among its own employees. Following the release of Sora, current and former OpenAI staff expressed misgivings. Boaz Barak, an OpenAI researcher and Harvard professor, noted that while Sora is 'technically amazing,' it's 'premature to congratulate ourselves on avoiding the pitfalls of other social media apps and deepfakes.' More remarkably, Josh Achiam, OpenAI's head of mission alignment, publicly questioned the company's trajectory, stating, 'We can’t be doing things that make us into a frightening power instead of a virtuous one. We have a duty to and a mission for all of humanity. The bar to pursue that duty is remarkably high.' Such statements from within the company highlight a profound crisis of conscience, suggesting that the fundamental belief in OpenAI's benevolent mission is faltering, even among those dedicated to it. This internal conflict may intensify as the company pursues artificial general intelligence.