The advent of AI has brought about a quiet lexical revolution. What do you call a chatbot when you have a "chat" with it? Is it really a conversation partner or just a machine? A programme? A piece of software? A chatbot is not a being, because it is not alive, but it does things in a way that doesn't necessarily feel machine-like. So, what should we call all of these AIs? Historian Yubal Harari has suggested the term "entity", which I don't dislike. It is in keeping with the dictionary definition, namely, "something that has separate and distinct existence and objective or conceptual reality". So far so good, right? Problems arise, though, when an AI is designated by referring to the job it performs. Let's not forget that, in English, a person who is stunning is a stunner, just like a piece of news that you find shocking can be referred to as "a shocker". Consider these sentences:
That was shocking news
It came as a shock to me
It shocked me to my core
That was quite a shocker
So, the "-er" ending does not imply humanity, ergo... a computer that teaches you something could technically be... a teacher? If we accept that label then we should also be ready to embrace the fact that an AI that books rooms for a hotel is an agent.😬
In the video below you can see a woman trying to get an AI to admit that it is not human and the AI refusing to accept that it is not a person. Just pay attention to the vocabulary that it uses.
Just like the rise of the cell phone made it necessary to coin the term "landline" to refer to the old telephones, and the invention of the electric guitar forced manufacturers to come up with the labels classical guitar or acoustic guitar, the popularity of AI-driven technology has brought about a series of changes which demand serious consideration.
The AI in the video self-identifies first as a "real authorized booking agent", then as a "live agent" and finally as a "virtual agent". Never as an AI, a language model, a computer program, a robot or a machine. I can only assume that very soon companies will try to normalize the label "virtual person" so they can say they have "people" answering phone calls without (technically) lying.
This brave new state of affairs is to me a reminder of the fact that some languages are better equipped than others to adapt to certain changes. In this particular case I think languages like Ojibwe(spoken in an area straddling the North of the USA and the south of Canada) have an interesting advantage over European tongues. Why? Because the grammar of Ojibwe considers a feature that Indo-European languages lack: animacy. To Ojibwe speakers nouns are animate or inanimate, that is to say, living or non-living. Interestingly enough, this feature can affect even verbs. For example, in order to say "the bird is white" an Ojibwe speaker must use the animate verb form waabishkizi, but the sentence "the shoe is white" uses the inanimate verb form: waabishkaa. To complicate matters further, sometimes a word can take the animate or inanimate form depending on the circumstances. Thus a doll, whose default form is inanimate, is referred to in the animate form when it is used in a ceremony. Long story short, it is a lot easier to hint at the to the nature of an AI in Ojibwe than it is in English or Spanish.
There is an online app called Friend that offers interactions with AI entities that behave like friends (!). And if you think that is creepy, don't look into Replika, which allows you to chat to an "AI companion" (or romantic partner). Is it okay to develop a relationship with a chatbot and then state that you have a friend?
It is becoming increasingly difficult to find the right words to name new realities. When I say I play the guitar I don't feel the need to specify if I refer to a classical, acoustic or electric guitar, but we might soon find ourselves in a situation in which words like agent or boyfriend might be ambiguous. Will we be compelled to refer to friends and partners without specifying if they are human or AI? Will it be acceptable to regard your interactions with a chatbot app as part of your "social life"? Will neologisms be necessary to name those connections? The very fact that I am considering these question already feels rather disturbing, but, then again, that is the world we're living in.
