TL;DR In the wake of the "Generative AI revolution", ChatGPT's initial popularity appears to be waning, with a nearly 10% drop in traffic in June. Interestingly, this dip coincides with summer school breaks, suggesting that a significant part of the tool's user base might be students using it for academic purposes. Amid these trends, it's necessary to reassess the overoptimistic expectations surrounding generative AI's broad application, as current data suggests a narrower immediate usage spectrum.

Some Recent Trends

As we stand on the cusp of what many have called the "Generative AI revolution", a surprising trend is unfolding – the initial spark of excitement around ChatGPT seems to be dimming, with traffic falling shortly after its boom in popularity. According to Similarweb, both mobile and desktop traffic to ChatGPT's website dropped by nearly 10% in June, marking a noticeable dip from its record-breaking growth trajectory.

For context Facebook exhibited consistent growth from their inception till now, although pace of growth has slowed in recent years.

To clarify not all GPT traffic has dropped, in fact developer interest has continued to grow with traffic to OpenAI's brand-new API platform, which empowers developers to utilize this technology in unique tools and apps, grew by 3.1% from May to June. This indicates that the developer community's interest in and usage of generative AI continues to grow.

Students are the key reason for ChatGPT's growth

Let's explore one possible reason for why this decline happened. There have been numerous use-cases suggested for AI Chatbots, ranging from personal assistants in healthcare, to "AI Attorneys". However this drop in traffic coincides with summer school breaks, suggesting an intriguing usage trend. Perhaps students using ChatGPT to write their papers isn't just one of the use-cases for the chatbot, it may be the main use-case.

Students using ChatGPT to write the their papers isn't just one of the use-cases for the chatbot, it may be the main use-case.

With this data, one can't help but question if the initial hype was a bit overly optimistic. Is the sudden deflation in ChatGPT's popularity an indication of the actual scope of its application? After all, about 90% of users are students leveraging it for academic papers.

Perhaps this may be one of the reasons OpenAI quietly shut down their "AI Detector" and only made a small mention in their original launch blog post. While one explanation offered is that the tool lacks full accuracy, this isn't a completely novel revelation. In fact, OpenAI had forewarned about the tool's "unreliability" during its initial launch. Another possible reason could be that, OpenAI may have realized its main growth driver is student essays and it isn't really interested in curbing usage of its core users.

‍

Looking forward

Silicon valley is well-known for its tendency for hyping up new technologies. Although the massive potential of generative AI cannot be denied, it's important to remain grounded. We at Rumi, expect growth and adoption of ChatGPT to grow but this data suggests that perhaps that in the immediate future there is a slightly narrower set of use-cases (or to be more precise - actual usage by users) than previously anticipated for generative AI chatbots. One thing to watch closely is how much ChatGPT traffic picks up once schools reopen.