Meta's No Language Left Behind model, spanning 200 languages including 55 African languages, forced small language AI startups worldwide to shut down after investors withdrew funding, AI safety researcher Timnit Gebru reported. "Facebook has solved it, so your little puny startup is not going to be able to do anything," investors told the companies, according to Gebru from the AI Now Institute.
The pattern repeats globally when OpenAI or Meta release large models. Investors in competing organizations across multiple continents "literally told them to close up shop," Gebru said. This consolidation threatens regional AI development from Lagos to Jakarta, where specialized models could serve local languages and contexts better than universal systems.
Universal models carry risks that impact users worldwide. Safety issues include fabricated medical transcriptions and undefined outputs in production systems. Multimodal large language models show cascading failures where struggles with one image analysis facet damage other aspects, researcher Javier Conde told IEEE Spectrum.
Gebru criticized Big Tech's approach: "People came along and decided that they want to build a machine god... they end up stealing data, killing the environment, exploiting labor in that process." The AI Now Institute's frugal AI research questions whether universal models represent optimal resource use for global populations.
The debate intensifies as enterprises worldwide adopt AI through automated ML systems and foundation models. Google DeepMind argues generative AI "unlocks general functionality" for robotics, unlike traditional robots trained on specific tasks. The global AI-powered humanoid robots market is projected to reach $7.73 billion as engineering advances.
Market consolidation eliminates diversity in international AI development. Organizations across Africa, Asia, and Latin America working on language-specific or specialized models face funding cuts when Silicon Valley announces competing universal systems. The fundamental question persists: whether one-size-fits-all models built in California serve users in Nairobi, Mumbai, or São Paulo better than locally-developed alternatives.