Aligned to whom?

The article 'Aligned to whom?' explores the complex and often ambiguous nature of AI alignment. It challenges the prevailing industry narrative that AI alignment is a purely technical problem focused on ensuring models follow human instructions. Instead, the author argues that alignment is fundamentally a political and economic issue, questioning whose values and interests are being prioritized during the development of powerful AI systems. The piece examines how corporate incentives, power dynamics, and societal biases shape the 'alignment' of models, often at the expense of broader public interest. By scrutinizing the goals of major AI labs, the author highlights the risks of allowing a small group of stakeholders to define the ethical parameters of technology that impacts global society. The article serves as a critical call to broaden the discourse on AI safety, moving beyond technical constraints to address the underlying power structures governing AI development.