From Pixels to Posts

Using social listening data to chart how generative AI could upend Instagram’s influencer economy

Influencers have transformed social media marketing, bringing creativity and connection. While challenges exist, their impact created new opportunities for brands to engage authentically with audiences. With influencers emerging across all sorts of niches from fashion to finance, advertising agencies use social listening tools to identify influencers in order to market goods and services.

We’ve written extensively about how generative AI and ChatGPT have upended industries like music composition, coding, and copywriting. Based on new social listening data and qualitative research, AI-generated content could also be coming for the social media influencer. Using large language models, agencies and brands can “generate” images of influencers created to sell specific products or market different services. We’ll look at this small but growing trend, analyze which platforms generated-influencers pop up, and showcase how social media users feel about this new change.

A small but growing trend

We first glance at the post volume to look at the underlying trend. You’ll note that there is a small volume within this space. However, we’re most interested in the growth rate and durability of the trend. You’ll note that across all platforms (but mainly Instagram), this trend’s growth emerged in April of 2023, well after the launch of ChatGPT, the watershed moment of democratizing generative AI. Since then, we’ve seen steady, durable growth within this trend. We fully expect this pattern to continue, as it shares the hallmarks of steady, consistent growth that we look for.

Figure 1: Post volume graph showcasing the growth of hashtag #AIInfluencer over the last two years (January 1, 2023 through present); Infegy data. 

An outsized share on Instagram

Keeping this small but mighty post volume in mind, we’ll examine which platforms AI influencer content appears on. Figure 2 shows 92% of all #AIInfluencer mentions on Instagram, a staggeringly high percentage. Additionally, while there have been several dips in concentration over the last two years, this concentration remains consistent.

The high concentration makes a lot of sense from our perspective. Despite efforts to pivot to a video-centric platform, photos within posts and stories remain very popular on Instagram. At its current stage, generative AI tools create realistic, reproducible human images with ease. However, creating AI-generated videos and audio is a bit more involved. As a result, we expect this concentration to continue until video-based generative AI technology becomes more widespread.

Figure 2: Channel distribution showcasing high Instagram penetration compared with other platforms (January 1, 2022 through present); Infegy data.

Conversations overwhelming female fashion-centered

Now that we’ve got a feel for which platforms #AIInfluencer content is popping up, let’s look at the topics and the demographics of who’s posting about it. We’ll jump to gender first. We examined other attributes about these posts, like source bios, to ensure that these accounts were not the AI influencers themselves.

We notice that 80% of people using the hashtag #AIInfluencer are women. Additionally, when women talk about AI influencers, they speak primarily on fashion-related topics. We see words like “dress,” “style,” “black,” and “model” appear repeatedly. Additionally, the growth around AI influencer content has come almost exclusively from women. Female posts have grown 604% percent within the last year, while those of men have only grown 140%.

This fashion focus showcases how AI influencers can upend the current influencer marketing sphere. Coordinating between many influencers for contracts, providing the products for placement, and managing the ongoing relationship and brand presentation can be prohibitive to many companies and brands. However, with fictitious influencers created by AI, these barriers no longer exist.

Figure 3: Gender distribution around people talking about AI influencers (January 1, 2023 through present); Infegy data.

How do consumers feel about this?

So far, we’ve left the most critical question to the end: How do customers feel about AI influencers? Professor Jan-Philipp Stein notes that customers seem not to care much and build similar parasocial relationships with AI influencers that they do with the real ones. We analyzed Emotions around how people discussed AI influencers (Figure 4) and came up with similar findings. We detected Love in 33%, Trust in 21% of posts, and Joy in 20% of posts. These metrics appear to be well in line with other traditional influencers we looked at.

Figure 4: Trends graph showing emotional distribution around AI influencers graph (January 1, 2023 through present); Infegy data.

Takeaways for your influencer identification process

Newly created AI influencers could massively disrupt the current influencer economy. We’ve shown a small, steadily increasing post count tagged with the term #AIInfluencer, and we expect this trend to grow over the next year. We’ve detected a high concentration within the female-dominated fashion-centered sphere on Instagram especially. Additionally, despite obvious ethical and economic concerns, consumer sentiments towards AI influencers mirror their emotions towards more traditional ones, with Love, Trust, and Joy predominant.

As practitioners in social listening, our key takeaways include recognizing the importance of monitoring this growing trend, especially on Instagram and adapting influencer identification processes to incorporate AI influencers. Brands should test the potential of AI-generated content to streamline marketing efforts, particularly in fashion-centric spheres. Additionally, the positive emotional responses from customers suggest an acceptance of AI influencers, encouraging brands to explore innovative strategies within this evolving landscape.