The huge surge of new generative AI users will slow

The launch of ChatGPT kick-started the use of genAI. The number of ChatGPT and genAI users will increase at a healthy clip over the next two years, but at a decreasing rate of growth. Even so, by 2025, over three-quarters (75.2%) of genAI users will use ChatGPT, and more than a third (34.0%) of the US population will use genAI.

GenAI use will increase across generations but will continue to skew young. The core age cohorts for AI use are 18 to 24 and 25 to 34. Both will hit around 54% penetration in 2025 (from a current 35.7% and 35.4% penetration, respectively).

To see the full forecast, click here.

• Men outnumber women in using genAI. Male genAI users currently outnumber female users roughly 2 to 1, but this will narrow somewhat by 2025. The differences relate, in part, to labor market participation, since most people using genAI do so for work-related activities.
• ChatGPT’s early lead will be challenged. Microsoft’s Bing, which uses OpenAI tech, and Google’s Bard launched soon after ChatGPT, but neither has yet offered a meaningful challenge. However, other programs—including Anthropic’s Claude, Baidu’s Ernie, Google’s next-generation chatbot based on its Gemini model, and many others—could threaten ChatGPT’s lead. We’ll know more by the end of H2.
• Open-source models, such as Meta’s LLaMA or Databricks’ Dolly 2.0, may accelerate growth. Meta has come down firmly in favor of open-source models, which it hopes will increase genAI development. These models are not quite as powerful as current closed models, such as those by Google and OpenAI, but they’re advancing quickly. By limiting development costs, open-source models will lower the bar for many smaller companies to build their own genAI models.
• The continued flow of investment capital should speed innovation. In H1 2023, global VC funding fell by almost half (48%), according to PitchBook, but genAI startups bucked the trend with investors committing more than $40 billion to AI startups during this period.
• Regulation is a wild card in our forecast. Algorithmic bias, copyright infringement, fraud by bad actors, and unease over the unpredictability of future developments have raised regulatory concerns. Most consumers and companies want some form of regulation, but there’s little consensus on the precise approach. That hasn’t stopped governments from trying, though. The EU, UK, China, and both the US Congress and the Biden administration are working on policies, all at various stages of development.

What does this mean for marketers? Unlike other technologies, genAI interfaces use existing language skills—a factor leading to broadscale adoption. Although Gen Zers and millennials will clearly be the most eager to embrace chatbots and other genAI tools, older consumers may get on board as well.