The cost of training AI models has surged post-OpenAI’s ChatGPT launch, but smaller firms find potential in AI. Stanford University’s report highlights tech giants’ dominance versus AI’s disruption for newcomers. Private-sector AI research intensifies, with the US leading in AI models. Despite soaring costs, AI investment in the US dwarfs China and the UK. AI’s transformative power could alter the global landscape, yet concerns linger over techno-feudalism and data colonialism by dominant players.
The cost of training AI models has risen significantly since OpenAI launched ChatGPT. Despite this, AI’s utilization presents smaller companies with opportunities to compete. A recent Artificial Intelligence Index Report from Stanford University illustrates contrasting stories: dominant tech giants such as Google, Meta, and Microsoft in research, investment, and AI model development, versus the disruptive potential of AI enabling smaller firms and new entrants.
Private-sector AI research has notably intensified. In 2011, 41% of new AI PhD researchers in North America entered academia; by 2022, this number plummeted to 20%, with 70% moving into industry roles. Consequently, the US leads with 61 notable AI models developed over the past two decades, compared to 25 from the EU and UK, and 15 from China. However, the cost for developing these models has soared, with OpenAI spending $78 million on GPT-4 and Google nearly $191 million on Gemini Ultra. In 2022, private-sector AI investment in the US reached $67.2 billion, dwarfing China’s $7.8 billion and the UK’s $3.8 billion.
Proponents argue AI could establish foundational infrastructure similar to 21st-century rail or telecom networks, potentially consolidating giant tech companies’ influence. Russell Wald, Deputy Director of Stanford’s Institute for Human-Centered Artificial Intelligence, emphasized the need for the public sector to maintain a stake in the AI landscape.
Emerging economies display greater enthusiasm for AI’s potential. Surveys indicate over 70% of respondents from Indonesia, Thailand, and Mexico view AI favorably, compared to 37% in the US and France. Emerging economies also report higher daily use of AI tools like ChatGPT. China, in particular, leads in practical AI applications, holding 61% of global AI patents and a significant share of industrial robot installations.
Demographics are key: 90% of the world’s youth reside outside the developed West and are eager to engage with the digital economy. Indian-born academic Payal Arora notes that for many in these countries, technology represents opportunity rather than a cause for pessimism.
Concerns persist that dominant US AI companies could create new realms of techno-feudalism or data colonialism, marginalizing emerging economies. Nonetheless, AI’s transformative potential may offer these regions a chance to alter the status quo.
