As we negotiate AI’s place in our society, it is important to consider how humanization contributes to issues with AI. Generative AI has the potential to do immense damage and good to our society, and in order to handle it responsibly we must understand it.
Recently, Albania appointed an AI Minister of public procurement: Diella. It is notable that Diella was presented to the public with a human name, avatar, and she/her pronouns.
This is common strategy where developers program their systems to be as human-like as possible, behaviorally and characteristically.
Humanization is used primarily to make generative AI easier and more enjoyable to communicate with. We see chatbots and virtual assistants generating exceptionally welcoming and friendly phrases.
OpenAI’s ChatGPT was asked, “How would ChatGPT greet a new user?” and said; “Hi there! Welcome to ChatGPT. I’m here to help you with just about anything you need — from answering questions and brainstorming ideas to helping with writing, learning, or solving problems. “Whether you’re just exploring or have something specific in mind, feel free to ask. There’s no pressure, and no such thing as a bad question here.”
ChatGPT’s response first engages the user, and then invites them to converse again with the algorithm. At the end of every following response, ChatGPT turns a question or command back to the user, initiating further interaction.
User interaction itself is what drives the expansion of the AI industry. Expansion generates profit, which is the corporation’s ultimate goal. By this logic, some would say that humanization is carried out in favor of economic growth, and is not of any real interest to the public.
Some would even argue that humanizing AI is dangerous above being unnecessary. Critics eagerly suggest short-term and long-term risks that the process could impose.
For example, that humanization drives job displacement within the humanitarian sector, or that it accompanies cybersecurity and exploitation risk.
Beyond economic ramifications, there are the interpersonal ones too. Heavy skepticism circulates around accountability concerning AI. When the algorithm generates something regrettable, the developers are reluctant to take responsibility for it. Gaps in accountability have historically been a slippery slope to corruption.
The systematic and interpersonal implications of humanization affect everyone. The practice is already well-established in the industry. So how can we avoid and reverse the negative impact of personifying AI?
The most feasible way to do so is through education. People, youth especially, should be properly informed on how generative AI works on a fundamental level. A society that aptly understands AI will apply it more effectively.
Public schools should implement full courses on the use of AI. Also, corporations should be informing the public about the risks that accompany/ come with using AI. People should interest themselves in the issue at hand, and hold themselves to a standard when handling AI.
Education and transparency will enable the population to make its proper judgement on humanization, and in return use AI honorably. With that, it would be a failure of the human race not to justfully do so.