“Made by a Human” is the new “Made in the USA”
90% of online content could be generated by AI by 2025.
“90% of online content could be generated by AI by 2025.” This is according to Nina Schick, a leading authority on AI and an advocate of responsible AI development, who made this prediction at the beginning of 2023.
As we observe the rapid proliferation of AI-generated content, it makes me wonder: will ‘Made by a Human’ be the new ‘Made in the USA’?
While human-made content often holds higher value than content generated by AI, the presence of both types of content is inevitable, and ultimately, can be a good thing. We will need and want both.
In light of what’s unfolding right before our eyes, I think it’s important for us to understand the unique value of human-made content...as well as to understand the value of AI-generated content. (Please note: This article does not cover the risks of AI-generated content, but you can find information on some of the risks here and here.)
Which Do You Prefer?
Made by Humans
Firstly, let’s clarify what “Made by Humans” means. In the context of this article, I am referring to it as content featuring real humans (like a video) as well as content orchestrated by humans (such as a script partially written by a human, rather than entirely by AI). This distinction is important as it acknowledges the value of AI as a tool in creative tasks, while also highlighting the importance of human involvement both behind the scenes and on-screen.
With the rise of Generative Artificial Intelligence and the content explosion that is still very much on the up-swing, my hypothesis is that content that is made by humans will hold higher value than AI-generated content because of two main factors:
1 — Supply-and-demand economics.
2 — Neuroscience and people watching.
Supply-and-Demand Economics
This is relatively straightforward — when there is less of something available AND when it holds value, then it’s value goes up; because there are more people who want it, and less of it available. This effect will be magnified in the future, as we find ourselves in a world where there is an increasing saturation of Generative AI-created content. In that context, human-made content will be relatively harder to find.
Human-made content will be harder to find.
Human-made content will be a relatively smaller piece of the pie and so those who are able to access it will be accessing something that may even be seen as somewhat luxurious. The social implications of having access to something that is rare will compound the value of human-made content, elevating it’s perception as well.
In addition to the perceived value of an increasingly rare set of content, it turns out that we have a hard-wired preference for, well, people watching. Let’s look at the science.
Neuroscience and People Watching
Studies of the human brain’s response to faces, body language and interpersonal interaction have shown that the brain is finely-tuned to process complex social cues from human interactions, as demonstrated by brain’s response to realistic, natural scenes depicting human interactions. These findings could imply that content with real humans engage the brain’s neural systems more effectively than interactions involving non-human agents, like avatars, potentially leading to a preference for human content. This is based on the idea that human interactions involve nuanced social cues that the brain has evolved to interpret, which might not be as richly or authentically presented by artificial agents.
We have higher levels of emotional and cognitive engagement when watching real humans.
The way that the human brain processes meaningful social interactions among real people influences neural processing in areas associated with face and body encoding. This research indicates that the human brain is sensitive to the realism and conformity of interactions between individuals, suggesting a sophisticated neural mechanism for decoding complex social information from real human interactions.
In another study, the focus was on the impact of meaningful versus meaningless interactions on brain activity. Here, the idea that the brain’s response to real, plausible human interactions is more nuanced and engaged compared to less realistic scenarios also supports a preference for seeing humans.
Real human interactions provide rich social cues that the brain is equipped to process.
Other studies look at Mirror Neurons and how the human Mirror Neuron System (MNS) responds better to interactive behaviors between humans compared to non-interactive behaviors. This suggests that the human brain is particularly tuned to understand and resonate with interactions involving real humans.
This supports the hypothesis that humans might have a preference for viewing content with real humans as opposed robots or avatars, as real human interactions engage the MNS more significantly. The MNS is believed to help in understanding others’ intentions, behaviors, and emotions, thereby playing a crucial role in social cognition.
More support for this hypothesis comes from a study done in 2023, aptly titled “Human Favoritism, Not AI Aversion” that found that when participants compared content created by humans vs that created by AI, revealing the source of the content (human or AI) influenced the perceived quality. Content identified as human-generated was rated higher, supporting the hypothesis that people are inclined to prefer human-created content.
Content identified as human-generated was rated higher.
However, the study makes clear that while there is a preference for human-generated content, there is no finding to support that humans have an aversion to AI-generated content. This bodes well for us and makes me hugely optimistic, because ultimately, I believe that we will need both types of content!
We Will Need Both
I’m heartened to come to the understanding that there is value to be found in both human and AI-generated content, because I believe that we will need both in order to evolve our society and bring more value to a larger part of the population. With that in mind, it’s helpful to understand when human-driven content is most effective versus when AI-created content is appropriate. Let’s focus our discussion on the psychological, emotional, and cognitive effects of consuming different types of content.
Humans — Content That Contains Real People
Real human interaction engages the Mirror Neuron System (MNS) more deeply, which facilitates a more nuanced understanding of complex emotions and intentions. This is important in contexts where empathy and deep emotional connections are key, such as in psychotherapy, complex learning environments, or intimate social interactions. Having access to content that contains real interactions provides the depth of emotional and behavioral cues that is needed for these types of situations (e.g. where high empathy and interpersonal engagement is necessary).
Another example of situations where access to content created by and containing real humans is where authenticity and trust are at a premium. This includes content like news and documentation because authenticity and trustworthiness are table-stakes. In these situations, watching real humans can provide a greater sense of credibility and reliability in news reporting, documentaries, and educational content, because not just understanding but also believing in the accuracy of the information are crucial to the success and value-perception of the content.
News and documentation require content to be created by humans because authenticity and trustworthiness are table-stakes.
Another example is when the value of something needs to be conveyed to the audience. Think of this as reading a product review. You would not trust a review written by a robot, you want to read reviews written by real people who really tried the product or experience they are writing about. Even better if that person is relatable to you, as their perception holds weight in your decision-making process. So here we find that the authenticity of real human expressions and interactions will build trust and lend credibility to the information being conveyed.
Another set of examples where human content is preferred is skill acquisition and complex learning. When it comes to the ways in which we learn hands-on skills, observing real humans performing the actions that we are trying to learn can be particularly beneficial. In this case, I’m referring to complex skill acquisition that requires understanding subtle nuances, such as in surgical procedures, mechanical repairs, or intricate craftsmanship. The precision and detail that we can observe in real humans doing these tasks will help us gain a deeper and more accurate understanding of performing these skills ourselves.
When it comes to the ways in which we learn hands-on skills, observing real humans performing the actions that we are trying to learn can be particularly beneficial.
Having given all these examples, let’s now look at examples of situations where content created by AI can be sufficient.
Generative AI — Avatars, Robots, and Cartoons
One example — which we have already seen for decades is in children’s learning and entertainment. Historically, this has not been Generative AI content, but it has been things like cartoons. This is effective because animated characters can simplify complex concepts and make learning more engaging for children. Animated characters and avatars can also be used effectively in video games and entertainment media to create engaging, imaginative, and fantastical experiences. This simplification of characters and their behaviors can help in reducing cognitive overload of the viewer, making learning more accessible and enjoyable.
Animated characters can simplify complex concepts.
Another example is in creative storytelling and games. Animation allows for exaggerated expressions and fantastical scenarios that are not bound by the physical rules of the real world, offering creative storytelling and visualization that can enhance imaginative thinking. With Generative AI providing the capabilities to produce many more storylines, game choices, scenarios, worlds, etc, this blows open the doors for creativity, allowing viewers to explore even more scenarios and concepts that go beyond real-world limitations. Creators can create more variations of their content, in the same amount of time, thereby expanding reach in a way that is beneficial not just for the creator but for the new audiences who get access and enjoyment where previously their needs and interests may not have been addressed.
Another example, and one that is somewhat more practical and has hugely positive implications for our society, is content generated for use in controlled environments. What do I mean by “controlled” environments? I mean like a laboratory, a curated interactive experience, a professional development workshop — anything where you have people who are coming in with a defined set of needs, expectations or goals.
A controlled environment with a defined set of expectations and goals is a good opportunity to use AI-Generated Content to facilitate positive outcomes at scale.
Therapeutic contexts, like therapy for PTSD, could use avatars (plus maybe virtual reality) to simulate situations. Training contexts, with company-specific content (e.g. continuing education for professionals), are good examples too because both repeatability and customization are required.
Being able to use Generative AI allows for the kind of scale and customization that will enhance the quality of the content and allow the content to be distributed more broadly, elevating society as a whole.
Using Generative AI for scale and customization in controlled use cases will elevate society as a whole.
Yet another example may be seen in practical commerce applications — being able to create an avatar of yourself, as the consumer, and put that avatar in situations or experiences or with products that help you as the viewer/consumer to make a purchase decision. Like, what if I want to create an avatar of myself to go clothes shopping? I would love to see what the clothes will look like on me (not on a model who has very little resemblance to me) and I would love to do that without having to go to a store and try thing on for hours. In a case like this, since the user is given agency and control in creating the avatar, there is less likelihood of that avatar being a deterrent in the same way as if it were an anonymous avatar with which the viewer could not connect or empathize.
“Made by a Human” is the new “Made in the USA”
Human-made content will always carry an intrinsic value due to its finite availability, authenticity and emotional depth; and this type of content will be valued at a premium. Having said that, AI-generated content will also play an increasingly valuable role over time.
Given that both can be valuable and both will continue to exist and proliferate, understanding when to prioritize human-driven content, and when AI-created material is sufficient, allows for a balanced approach to content creation and consumption.
We need to understand when to prioritize human-driven content, and when AI-created material is sufficient.
Being aware of this gives us the opportunity to lean in to the unique advantages of both human creativity and AI efficiency. Ultimately, we can enhance learning, entertainment, and practical applications, enrich society and expand our collective capabilities.
Read More About Generative AI
Subscribe to our newsletter at TechExplorer.AI.
Read about how to tell whether something is created by AI (it’s not as simple as we wish).
References
Nicola Canessa, Federica Alemanno, Federica Riva, Alberto Zani, Alice Mado Proverbio, Nicola Mannara, Daniela Perani, Stefano F. Cappa, The Neural Bases of Social Intention Understanding: The Role of Interaction Goals, PLOS ONE (2012). Available at: https://doi.org/10.1371/journal.pone.0042347. (https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0042347)
Susanne Quadflieg, Francesco Gentile, Bruno Rossion,
The neural basis of perceiving person interactions,
Cortex (2015). Available at: https://doi.org/10.1016/j.cortex.2014.12.020.
(https://www.sciencedirect.com/science/article/pii/S0010945215000295)
Tsuyoshi Aihara, Shinji Yamamoto, Hirotaka Mori, Keisuke Kushiro, Shintaro Uehara, Observation of interactive behavior increases corticospinal excitability in humans: A transcranial magnetic stimulation study, Brain and Cognition (2015). Available at: https://doi.org/10.1016/j.bandc.2015.09.003.
(https://www.sciencedirect.com/science/article/pii/S0278262615300208)
Galit Yovel, Neural and cognitive face-selective markers: An integrative review, Neuropsychologia (2016). Available at: https://doi.org/10.1016/j.neuropsychologia.2015.09.026.
(https://www.sciencedirect.com/science/article/pii/S0028393215301688)
Zhang, Yunhao and Gosline, Renee, Human Favoritism, Not AI Aversion: People’s Perceptions (and Bias) Toward Generative AI, Human Experts, and Human-GAI Collaboration in Persuasive Content Generation, MIT (May 20, 2023). Available at SSRN: https://ssrn.com/abstract=4453958 or http://dx.doi.org/10.2139/ssrn.4453958
