In a growing chorus of voices, prominent figures in the tech industry, including Jimmy Wales, Tim Berners Lee, and Signal President Meredith Whittaker, are raising red flags about the alarming concentration of power in the development of artificial intelligence (AI). Concerns center around the potential ramifications of a handful of companies dominating AI, leading to unprecedented control over evolving technology and, consequently, society.
The catalyst for this discussion was the surge of interest in OpenAI's ChatGPT, a chatbot that captured attention last year with its unique ability to respond to user prompts. This innovation ignited what industry insiders are calling an "AI arms race," as major players like Microsoft and Google aggressively pursue the creation and deployment of their AI models, demanding colossal computing power and vast datasets for training.
Whittaker, the President of encrypted messaging app Signal, emphasized the worrisome concentration of AI development resources, stating, "Right now, there are only a handful of companies with the resources needed to create these large-scale AI models and deploy them at scale. And we need to recognize that this is giving them inordinate power over our lives and institutions."
This sentiment is not isolated, with former Los Angeles Dodgers owner Frank McCourt, now heading Project Liberty, echoing concerns about the overwhelming influence held by a select few. He cautioned that if changes are not made, a mere handful of platforms will control the AI landscape, leaving little room for competition.
A common thread in these warnings is the assertion that AI, as it stands, is an outgrowth of centralized corporate power. Whittaker argued that the current state of AI is built on the accumulated resources of large tech corporations, primarily based in the U.S. and China. These corporations, she claims, gained dominance through the surveillance advertising business model, providing them with robust computational infrastructure, extensive datasets, and market dominance.
McCourt and Whittaker argue that users have lost control over their data online, and tech giants exploit this data for profit. Project Liberty's manifesto states, "Big tech and social media giants are inflicting profound damage on our society," and McCourt fears that AI could exacerbate this issue, labeling generative AI as "centralized, autocratic surveillance technology."
While industry veterans express concerns about the concentration of AI power, Jimmy Wales sees potential disruption. He points to leaked Google memos indicating that open-source models, not owned by single entities like Google or Microsoft, pose a threat to established giants. Wales suggests that open-source models, developed and expanded by anyone, could democratize AI and pave the way for diverse, accessible applications.
As the debate over the democratization of AI intensifies, industry leaders grapple with the implications of a future where a select few wield immense control over technology that increasingly shapes our lives. The question remains: Can AI evolve in a way that preserves democratic values, or will the current trajectory further consolidate power in the hands of a privileged few?