Leaders Opinion: A Year of Stability AI’s Trials and Errors

San Francisco-based AI firm, Stability AI, celebrated its one-year anniversary in the generative AI space. CEO Emad Mostaque expressed pride in their open-source AI models, with aspirations of a Nobel Prize for their contributions. Stability AI's models are highly popular, claiming an 81.32% market share, surpassing competitors like Midjourney and Adobe.

To conclude our exploration of this topic, we sought the perspective of Manish Sethia, Vice President - Applied AI at Chubb. Manish Sethia said, 'Stability AI has just completed its first year, but it has already made a splash in the GenAI Image space with its Stable Diffusion product and more recently, with the release of its interactive counterpart, DreamStudio.

The growing popularity of AI art generation tools but also the legal and ethical issues they raise, such as concerns over privacy, misinformation, and lack of context. Lensa, a problematic application rooted in diffusion models, drew attention for generating avatars that emphasized certain body features or altered facial characteristics.

Manish pointed out that things are not as rosy as they seem for the company. Allegations of unpaid invoices and questions about the CEO’s degree credentials do not bode well for the company’s image. Increasing concerns about privacy breaches, copyright infringement, racial and gender bias, and other issues are raising ethical and legal concerns about AI-generated images overall, further complicating the landscape. Companies like Google DeepMind are paving the way to deal ethically with some of these issues and could be the path for Stability AI to follow to establish itself as a long-term major player in the space.