AI is getting very good at making things. It can write code, generate images, draft documents, build presentations, create music, and apparently produce more than 50 emotes for a streaming channel in about three minutes. Depending on who you ask, that last one is either a pretty neat use of technology or evidence that civilization is circling the drain.
The tweet that kicked off what I’ve started thinking of as "AIEmoteGate" was simple enough. A streamer showed off a set of emotes he had generated with ChatGPT, and the reaction quickly turned into a much bigger argument about AI, creativity, and the value of human work.
What caught my attention was what people were really arguing about: what counts as meaningful human contribution once AI can handle more of the production. That question has been showing up in other places too, from corporate investments in "human skills" to my own experience building software with AI, and it feels like it is going to matter a lot more as these tools get better.
When the Easy Part Gets Easier
 |
| The eye still matters. |
That struck me because those are exactly the kinds of skills that become more important when AI starts taking care of more of the routine execution. If producing the first draft, running the analysis, generating the image, or writing the basic code takes less time, then more of the value naturally shifts toward deciding what should happen next. Was the analysis asking the right question? Does the image fit the audience? Is the code solving the right problem? Is the recommendation actually useful?
We tend to talk about AI productivity in terms of how much faster someone can get work done, and that absolutely matters. But as the mechanics get easier, simply being good at the mechanics becomes less of a differentiator. As that happens, companies are going to care more about the things AI still doesn’t do particularly well on its own.
EY’s announcement is interesting because they are attaching real money to the idea that some human skills may become more valuable as AI takes on more of the routine work. Judgment, adaptability, and business acumen are being treated as core parts of how the company expects people to work alongside AI. That makes the investment feel like a pretty clear signal about what EY thinks is going to matter more in the future.
What Happens When the Tool Gets Out of the Way
I have seen a smaller version of this play out in my own work. Earlier this year I spent a few sessions using AI to help me build Stream Deck plugins, something I wrote about in When Software Stops Being the Bottleneck. What surprised me was how quickly the coding stopped being the thing that determined whether an idea was worth pursuing.
That changed the way I worked. Instead of spending most of my energy figuring out how to implement something, I could spend more of it deciding what I actually wanted the plugin to do, whether a suggestion made sense, and whether the result was useful enough to keep. AI made the technical part dramatically easier, but it did not decide which ideas were worth building or when the implementation had gone off course.
(My experience lines up with broader research on human-AI collaboration, which has found that the way people work with AI can matter as much as simply having access to it.)
That is the part of AI collaboration that gets lost when we reduce the conversation to "AI did the work." In my case, I was not handing the problem over and waiting for a finished answer. I was using AI to make the process easier, but it still depended heavily on my own judgment. The software became easier to produce, which made the decisions around the software more important.
In that sense, the builder was still very much involved, but their time was being spent differently.
When the Thing Being Made Isn’t the Whole Thing
That brings me back to the emotes.
The original post was pretty straightforward. An X user, Dirk_JDR, said he gave ChatGPT a prompt, got more than 50 emotes in about three minutes, and used them to rebrand his Kick channel. That was enough to set off a much bigger argument about AI, artists, creativity, and whether using AI-generated assets says something about the creator himself.
What interested me was how much weight people were putting on the emotes. Emotes matter on a streaming channel. They help shape the community, they become inside jokes, and good ones can absolutely add personality. But they are still supporting assets. People do not usually decide to spend hours watching a streamer because the emote package is especially impressive. They stay because of the person on the other side of the screen.
That is where I think some of the argument gets tangled. If AI lowers the cost of producing a supporting asset, that does not automatically lower the value of the person using it. The streamer still has to be entertaining. The musician still has to make music people care about. The presenter still has to have something useful to say. The existence of cheaper production tools changes part of the process, but it does not necessarily change what the audience ultimately values.
There is still a legitimate argument about what happens to the artist who might otherwise have been paid to create those emotes. That matters, but it is a different question. What I find less convincing is the leap from "AI made this asset" to "therefore the creator has less value." Those two things are not the same.
When Everyone Can Make Something
One of the strangest things about AI is that it can give you too many choices almost as easily as it can give you one. Ask for an image and you can get variations. Ask for a paragraph and you can get five rewrites. Ask for a solution to a coding problem and you may get several approaches that all look plausible. The bottleneck starts moving from creation to selection.
That sounds like a small change, but I think it matters. If almost anyone can generate a decent first pass, then the first pass itself becomes less impressive. The harder part becomes knowing which version is actually right for the situation, which one fits the audience, which one feels like you, and which one should be thrown away even though it looks perfectly competent. Researchers studying human-AI interaction have described a similar shift, where automation moves people from producing the work to evaluating it.
This is where taste becomes more than a creative luxury, i.e. it becomes part of the work. Knowing which option fits the situation becomes its own skill, and someone still has to decide what survives. If producing things gets easier for everyone, this may be one of the places where people start to stand out. As making things gets easier, knowing what is worth keeping starts to matter a lot more.
There is Still a Line
None of this means every use of AI is interchangeable or harmless. There is a meaningful difference between using AI to help create something and using it to replace the person whose identity, judgment, or voice is the reason people cared in the first place.
That distinction matters more in some areas than others. Using AI to help build a plugin or create a set of emotes still leaves the person responsible for the idea, the choices, and the finished result. The line starts to move when AI begins speaking for the person, making decisions in their place, or becoming the personality the audience is supposedly connecting with.
That is where questions about authenticity become much harder. People are usually willing to accept tools that help someone express an idea more effectively. They are much less comfortable when the tool begins to substitute for the person behind the idea. The boundary will not always be obvious, and different people will draw it in different places. But I think it matters because collaboration can make someone more capable, while substitution can eventually start removing the person from the equation.
So What Are Humans for?
We have been through this kind of change before. When industrial robots started showing up on assembly lines, the wailing and gnashing of teeth was louder than a KISS concert. The fear was easy to understand: if machines could do the work faster, more consistently, and without getting tired, what would happen to the people doing that work?
What happened instead was adaptation. The work changed around the machines. People moved toward operating them, maintaining them, improving the process, handling exceptions, and doing the things the machines were not good at. Assembly got faster. Quality got better. Human contribution did not disappear. It moved.
AI may be the digital version of that same transition.
If generating the first draft, the image, the code, the presentation, or the supporting asset becomes cheap and easy, then those things stop being the only places where value lives. More of the value starts showing up in the choices around the work: what to make, what to change, what to keep, and what is actually worth doing.
So if AI can make everything, what are humans for? Probably for the same reason we have always mattered: deciding what is worth making, what is worth keeping, and what actually matters.