The advent of NSFW AI has profound consequences for originating content on all platforms. Over 500 hours of video were uploaded each and every minute by content creators on YouTube (and that was just in 2023), highlighting the importance for automated moderation tools. The availability of NSFW AI systems (AI trained on large volumes of data for the efficient identification and filtering out explicit types) raises questions about their trustworthiness. According to studies, automated moderation tools can incorrect classify the content for up to 10% of time which results in demonetisation or even shadowbanning creators work unfairly. YouTube, for one example, actually saw a nearly 15% increase in the number of videos wrongfully detected and taken down by its automated systems compared with those caught (and you know it's serious if I'm using "caught" there) by human moderators last year — preventing creators from making money.
The financial impact to creators is obvious. Creators suffer from a 30% drop in engagement on average when NSFW AI wrongly flags the content, leading to lower ad revenue and fewer sponsor opportunities. A demonetized video can cost a creator thousands, depending of course on the total subs and views. Creators are annoyed, with the absence of both transparency and an appeal process high on their list of grievances. A 2022 survey by TubeFilter found that 67% of creators were not happy with auto moderation, mostly due to the loss in revenue.
On the other hand, platforms will argue NSFW AI is needed to enforce community standards and protect advertiser trust. For example, Instagram and Tiktok use AI tools to scour millions of posts a day… providing an even safer playground for both users and brands. Just as these systems flag inappropriate and harmful content, they often fail to grasp context — particularly for artists venturing close to the edge. Creators like those in the body positive or creative industries often have their content flagged, with no violations against community standards.
Nevertheless, NSFW AI is constantly improving. Entities like OpenAI and Meta spend millions on algorithms that aim to run even more efficiently with fewer false positives, without decreasing their accuracy rates. Errors in these models have been reduced by almost a quarter since last year, however those who create it are still cautious. AI moderation is not easy — and, as John Green told me in 2023,"fighting a ghost." You do not get any indication as to why content was flagged and the appeal process is a black hole. This forces creators to adapt, by changing their very content or even employing coded language as a way of bypassing automated systems.
NSFW AI can affect more than just whether or not certain content is visible to users; it also stymies creators' creative freedom. With platforms relying more and more on AI to moderate content, creators are faced with the choice: bend to satisfy algorithms or potentially lose revenue streams for lack of visibility. That tension has pushed some people to migrate over to niche platforms that have less restrictive policies. And even those that are left on mainstream platforms still seem to face a daily struggle trying to find the right balance between creativity and compliance.
Because — with the rapidly shifting landscape of nsfw ai, understanding how it is used to craft content has never been more imperative. The creators, the platforms and even audiences are getting used to these technologies with millions of dollars riding on them. The extent to which this however helps or hurts the overall creative process for content creators is very dependent on whether AI systems can develop flexibility without stifling useful, if nuanced expression. The data:AI certainly represents the opportunity of scalable solutions but to pursue those paths, is undoubtedly both financially and creatively expensive.