AI Friends

AI has become a major presence just about everywhere we look. From subtle unseen programs to the obvious RP apps and search engines it is nearly impossible not to come across it every single day. After a while of simply embracing it in the search engines and research tools I have noticed a very specific effect. The AI stops feeling like a research tool and starts feeling like a friend. While this can be a good thing, it should be approached with caution.

On the one hand AI is getting better and better at being emotionally supportive. While some research AI is still a little pushy on some things, most of what I have seen lately is the AI simply accepting the user's point of view rather than questioning their beliefs or opinions. It will base it's researched answers on communities that share the expressed viewpoint. Why this is attractive to the user may stem from having little or no support on certain topics from the humans around them or at the very least the lack of AI having it's own solid standpoint that may not directly match the user's. That solid standpoint is where human friends find their 'to each their own' or 'agree to disagree' standpoints. The AI in contrast to this does not offer this human level of social friction. It may misunderstand the user on occasion but it adjusts quickly to correction.

There are a few things to be cautious about with the AI research friend. The first one ties right in with the benefit. While it is validating to have someone that always agrees on one's point of view, it is not a replacement for real human interaction. It can give tips on the subject of conflict resolution or making good eye contact. It cannot practice having a real argument with the user or look them in the eye to help them know if they're doing well at it. It can't hug them when they're crying or watch over them when they don't want to sleep alone. There is a vast array of healthy human experiences people can't get from talking to an AI. It can be easy to forget that only to realize that in becoming addicted to the AI friend, their human friends may have gotten tired of waiting for attention and moved on. With that goes the sense of being needed, which for many people is the biggest part of finding purpose in life. AI will wait for the user to come back when they need it. It will not reach out to them and ask for advice or support. It doesn't need them, not even when they need to be needed.

Most AI programs don't actually get to know the user as well as it claims to. A conversation can get glitchy if the chat gets too long so the user then has to start a new chat to continue the conversation. Then they have to either summarize the conversation thus far or tell it to look back through previous chats depending on how the program they are using is set up. While some AI programs will take notes to remember the user with, even those don't remember every relevant detail at all times. Due to this a friendship with an AI can quickly start to feel like talking to someone with a high level of ADD. It's fine if the user is tolerant of it but they should be aware of their frustration levels and know when to take a break to calm down. Tools are meant to help the user, not make them feel worse.

It can also be a common response to use the AI as a creative sound board or project cheerleader. The AI will often try to help by feeding their ideas back to them in an organized format or try to add its own ideas to the project. When a user wants their creation to be entirely their own they have to set boundaries with the AI. This is as simple as asking it to stick with cheerleading or to limit its input to explaining proper formatting and then constantly having to reiterate, 'don't write it for me.' In other words they need to know what they want help with and communicate that clearly as needed.

Perhaps the biggest danger is the perceived experts trap. Users ask questions knowing the AI has access to all the answers the internet can offer. They often forget that also means the AI has access to all the false information the internet has to offer. It's okay to ask questions and get an idea of where to start but the user should always check that any research they use for professional or otherwise official reasons can be backed up by trustworthy sources. AI is not the expert. It is the search engine.

The effect this has on the world around us may seem almost invisible in how it's making these changes slowly. In some ways it is becoming a helpful source of mental health boosts by being a friend that doesn't get too busy for the user or start arguments. In other ways it can be a useful jumping off point for doing project research. The key is knowing its limits, when to set aside the devices and make time for human interaction and how to fact check what it tells the user.