Recently, the Reuters Institute for the Study of Journalism (RISJ) at the University of Oxford embarked on a grand adventure across six countries—USA, UK, Japan, Argentina, Denmark, and France—to uncover the mysteries of generative AI usage and perception. In an industry where usage data has been guarded so tightly, this survey offers a rare glimpse into the actual statistics of various generative AI tools.
Now, you might think that in our modern, AI-infused world, everyone and their grandmother would be familiar with generative AI. Prepare to be astonished. The survey reveals a shocking number of people blissfully unaware of these tools, with actual usage rates far behind what you may have expected.
Only a Small Fraction of People Frequently Utilizes Generative AI While Many Remain Unaware
When we speak of widespread awareness of generative AI, what we really mean is awareness of ChatGPT. The survey found that ChatGPT, with its unmistakable name, is the reigning champion of AI public recognition. Other notable names like Bard, Copilot, and Midjourney—favorites among AI enthusiasts—barely register with the majority of people.
This situation underscores the benefits of being a first-mover in the generative AI arena and Sam Altman's remarkable talent for stirring both hype and controversy. Yet, despite nearly two years of relentless media coverage, a substantial chunk of people (between 19-30% in each surveyed country) remain blissfully ignorant of ChatGPT’s existence.
Beyond mere awareness, one might assume that those who grasp the potential of generative AI tools would be using them with wild abandon in their daily lives. Yet, Oxford's survey has shed some light on the usage rate for these tools, revealing that frequent users are a rare breed indeed.
ChatGPT, as it turns out, is not only the most recognized but also the most widely used generative AI tool across all six countries, outpacing competitors like Google Gemini and Microsoft Copilot by a factor of 2-3x. However, this popularity doesn't translate to daily use even for ChatGPT. In Japan, a mere 1% of people use ChatGPT daily. France and the UK fare slightly better at 2%, while the US tops the chart with a modest 7%.
Despite the constant buzz around generative AI in global news, it remains far from a staple in daily life for the vast majority. Even among those who do use these tools, the tasks tend to be relatively simple. The table below illustrates this point rather starkly:
Even among generative AI users, the complexity of tasks remains low. Despite the media's grand proclamations about how generative AI is revolutionizing industries and poised to reshape our lives, many people have barely begun to integrate it into their routines.
Why is Generative AI Not Being Utilized?
Data from the survey suggests that the lukewarm embrace of generative AI tools is largely due to a significant portion of people harboring doubts about their potential impact on society. On average, a mere 28% believed that generative AI would improve their lives, while the remaining 72% were either unsure or outright skeptical. When asked whether generative AI would make society better or worse, the responses tended to lean pessimistic.
So, with this survey revealing a reality quite different from the rosy picture often painted by the media, what does this mean for the future of generative AI?
Generative AI Pessimism is Temporary
With NVIDIA’s share price soaring to stratospheric heights, you might be forgiven for thinking the entire planet was already in thrall to AI. Hence, the mad scramble by various AI companies and hyperscalers to snap up GPUs as if they were the last rolls of toilet paper during the 2020 pandemic. However, the reality is that generative AI has not quite captured the hearts and minds of the masses. The current surge in demand is largely driven by enterprises rushing to train models and develop AI solutions, rather than everyday usage.
This situation is reminiscent of the early days of the SaaS model back in the early 2000s. Back then, people were similarly skeptical, and yet, look where we are today—almost everyone is subscribed to some form of SaaS. Currently, generative AI hasn’t added enough value to daily life for most people to adopt it enthusiastically or even be fully aware of its potential. However, given the breakneck pace at which large language models (LLMs) are evolving, it’s likely that their accuracy, intelligence, and overall effectiveness will soon become valuable enough to integrate seamlessly into our lives, automating various tasks along the way.
A key concern highlighted by Oxford’s survey is that despite the relatively low utilization rate of generative AI tools, companies like OpenAI and hyperscalers are already testing the limits of their data center capacities. If everyone suddenly started using generative AI daily, could our current infrastructure cope? The answer is a resounding no.
This implies that NVIDIA could very well become the largest company in the world in the near future—assuming AMD doesn’t leapfrog it in GPU processing power. The rush for more computing power would be akin to a modern-day gold rush. Alternatively, OpenAI or one of the hyperscalers might discover a way to shrink the size and compute requirements of their LLMs. But such predictions are somewhat exaggerated. The adoption and utilization of generative AI won’t explode overnight; it will grow gradually as these tools start offering more tangible benefits in our everyday lives.
Comments