Understanding Anthropomorphism: From Ancient Beliefs to AI Systems like ChatGPT
One of my ADHD superpowers is that when I get hooked on a topic, I go all in. It’s the kind of obsession that keeps me up all night, scouring the internet for every last nugget of information. Last summer, my obsession was Eminem. I dove headfirst into every album, watched every interview, tuned into his daughter’s podcast, and even learned about each of his music-industry feuds.
Lately, though, I’ve been completely absorbed in the world of AI. With something new happening every day and the constant release of new tools, it’s like an all-you-can-eat buffet for my curiosity. Something else that fuels my dopamine-seeking brain? The philosophical questions around AI. I love thought-provoking conversations about what it means to be human and why we create machines that mimic us. And machines are getting really good at mimicking us- so good that even the brightest of tech minds have been left wondering if bots are sentient. This is where the term “anthropomorphism” comes into play. In this post, I want to dive into what anthropomorphism is, why we do it, and why it’s important we recognize our tendency to anthropomorphize AI.
What is Anthropomorphism?
Have you ever assigned human characteristics to your pet, your car, or even a plant? Congratulations, you’ve engaged in anthropomorphism. As humans, we naturally tend to anthropomorphize, which means attributing human traits to non-human things. It’s something we’ve been doing for centuries to help make sense of the world around us. In ancient civilizations, natural elements were given human characteristics to help people comprehend the world. Today, we see it everywhere- in literature, art, and even everyday language. Just look at any Disney movie and you’ll see all kinds of animals with human traits (and sometimes even singing!).
Why We Anthropomorphize
So why do we turn our pets, cars, and tech into little humans? Researchers think our tendency to anthropomorphize can be explained from an evolutionary perspective. Anthropomorphism can be a useful tool for us when we deal with complex agents like animals or machines as it helps us predict their behavior. But it can also be dangerous, especially when talking about something like an animal from a scientific perspective, to assume that they have the same emotions or cognitive experiences as we do. Anthropomorphizing also makes us feel more connected to our surroundings. In fact, studies suggest that just using the term “Mother Nature” increased people’s sense of connectedness to nature and helped them act more eco-friendly. Socially, it helps us to feel a sense of belonging. Although anthropomorphism can be a useful tool, it’s important to remember that we’re projecting our own thoughts and feelings onto something that isn’t human.
Anthropomorphism and AI Design
In addition to anthropomorphizing animals or natural forces, we also create human-like machines because of how we understand human cognition. Neural networks, for example, are inspired by the brain’s structure and function. Also, people prefer to interact with NLP models like chatGPT using natural language because it feels more comfortable. Creating useful AI systems happens when AI designers are able to replicate human qualities like empathy and vulnerability. And, in some cases, ChatGPT might be even better at empathy than humans. In a recent study, ChatGPT outperformed human physicians by 41% when it came to displaying empathy. Research shows that people respond more positively to AI systems that have human-like qualities like displaying emotions, showing empathy, or using natural language. Users prefer to feel a psychological closeness to the AI, as demonstrated in a study where participants were more motivated to increase charitable giving to nonprofit AI agents that looked and smiled like humans. By building AI that understands and mimics human emotions, we work towards better communication between humans and machines.
Dangers of anthropomorphism
While anthropomorphism can absolutely be a tool, we need to consider the ethical implications of creating these human-like AI systems. In order to create a system that displays empathy, it needs to be trained on a large amount of data. It’s important that we hold AI developers accountable when it comes to training data and make sure sensitive information isn’t misused as AI systems can be trained off of personal data from users.
As AI systems become even more human-like, they will understand human emotions even more and could potentially manipulate human emotions for profit or other sinister purposes. Luckily most organizations have guidelines in place for responsible development to be respectful of human rights and ethical in training their AI models.
Outside of these ethical concerns, we also risk overestimating AI capabilities or misunderstanding AI behavior by anthropomorphizing AI systems.
Limitations in Creating Human-Like AI
Current AI systems are getting more and more impressive in their ability to understand and respond to human emotions. That said, they still have limitations. They can’t always understand complex emotional states or display empathy in a way that’s authentic. The development of AI systems that are human-like requires collaboration and input from other fields like psychology, neurosciences, and social science. As we move forward in creating human-like machines, I think there are a few questions we need to be asking ourselves:
- How can we make sure that these AI systems don’t manipulate human emotions?
- How can we prevent the misuse of personal data when training these systems?
- To what extent should AI systems replicate human emotions and traits?
Wrapping Up: Anthropomorphism and AI
Anthropomorphism is a deeply ingrained human tendency to give human traits to non-human things. For centuries, this has shaped our interactions with our environment, animals, objects, and now AI systems. Incorporating human-like qualities in AI systems helps us to build more meaningful connections between humans and machines. Anthropomorphism also has dangers such as privacy concerns, the potential for emotional manipulation, and risks of unethical development. As we continue to create advanced AI systems, it’s extremely important that we’re aware of both the potential benefits and risks of anthropomorphizing this rapidly evolving technology.
Thanks for reading! For more, you can follow me on Twitter @brittanynpotter or contact me at workwithpotter@gmail.com.
