It’s time to take human-AI communication seriously once more. The advent of sophisticated chatbots like OpenAI’s ChatGPT has made effective human-AI communication amusing and widely available. Even to pupils who write essays, which has given rise to a few academic worries.
When reading something, how will we know whether it was authored by an AI, and why is that significant? When we leave a remark on an essay or article, to whom are we responding? We might explore how we might employ these new chatbots in our learning by reframing the topic in light of the philosophical history of dialogue.
More advanced AIs
There is a long history between dialogue and AI. In order to determine if an AI was intelligent, English mathematician Alan Turing developed a “imitation game” in which a judge typed a conversation with one that was claiming to be a human.
What may AI chatbots of today have to say about such a conversation? Lex.page is a writing tool with AI capabilities that can come up with material or provide answers. When I questioned the Lex.page AI about dialogue and AI, it responded as follows:
AI and discourse are related in a number of different ways. These can be used to have dialogues and interactions with users that are authentic and human-like, allowing computers and other devices to comprehend and react appropriately to user inputs. Additionally, it can be used to examine and understand conversational data, enhancing customer support, identifying fraud, and more.
The Turing Test’s attraction was that it avoided explicitly defining intelligence and that it relied on conversation, one of the most distinctly human of interactions. According to Lex.page AI’s description:
A human would converse with two entities in this so-called “Turing Test,” one of which is a machine, without being able to tell which one is the machine. The AI has passed the Turing Test if a human cannot detect the difference between it and the other creature.
For many years, communication between humans and computers took the form of a command line, where you could command an operating system. As an illustration, consider Joseph Weizenbaum’s ELIZA, which interacted with you like a therapist and questioned your thoughts on whatever you mentioned.
The creation of Large Language Models (LLMs), which are trained on trillions of pages primarily scraped from the web, has changed this. These are far more literate and are able to carry on a discussion or even come up with brief articles on various subjects.
The Turing Test was a wonderful tool to determine whether an AI-driven system could deceive people by acting like them. Sundar Pichai, the CEO of Google, announced Duplex in 2018, a voice assistant that could schedule a hair appointment without indicating that it was an AI.
The fact that a conversation with the Language Model for Dialogue Applications (LaMDA) persuaded Google engineer Blake Lemoine that the AI might be sentient and hence merit ethical concern is not surprising.
If I didn’t know exactly what it was, Lemoine remarked, “I’d…
I’d guess it was a 7 or 8-year-old child who also happens to be knowledgeable in physics. When he presented the transcripts to superiors, they disregarded the evidence, and after Lemoine made his ethical concerns public, he was put on paid leave.
Then what? Perhaps we can review previous philosophical discussions about dialogue.
Discussion of philosophy
Philosophy has a long history of discussing challenging subjects in conversation. A form of writing known as dialogue can serve as a model for education, inquiry, and wise conversation.
Socrates is depicted as practising philosophy through dialogue in Plato and Xenophon’s dialogues. Plato and Xenophon explained the purposes of dialogue and provided models that we might still learn from 2000 years later by asking questions and thinking about the answers.
In my work, Defining Dialogue, I describe how the popularity of the writing genre of dialogue fluctuates as the culture of inquiry shifts. Theorists like Mikhail Bakhtin have more recently conceptualised this method of engagement.
Dialogic writing was favoured during the period of Plato and Xenophon as a manner of philosophical expression. The exception in later times was David Hume’s Dialogues Concerning Natural Religion (1779). They were created to deal with sensitive topics on which an author might want to avoid taking a firm stance.
Plato contrasts dialogue stanzas with scripted speeches in the Phaedrus. He portrays Socrates as the master orator who was then promoting dialogue as being superior. A speech cannot adjust to a listener or reader like written articles may. Contrarily, dialogue engages listeners in a way that AI chatbots may also be developed to do.
Additionally, as Lex.page noted
In Xenophon’s depiction, Socrates would pose a series of queries to elicit the thoughts of his interlocutor, frequently turning the dialogue around to get an opposing viewpoint in order to more thoroughly consider the argument. He would also perform dialectic, which is the process of examining ideas in search of the truth.
Logical conversation with machines
The time for dialogue has once again arrived with the development of chatbots. I contend that we can benefit from the accessibility of these chatty devices.
You may interact with the ethics professor I made, for instance, by utilising Character.AI.
With the help of AI, you can build a fictitious persona with whom you can communicate.
Users can interrogate the professor (or other characters) to record a dialogue, which is impossible with a standard textbook. They shouldn’t, however, believe all the lecturer says. Being the character. Everything the characters say is made up, according to AI site. You might be able to persuade it to acknowledge that it is unethical to try to deceive us by acting like a human, something I was unable to do.
I encourage my students to use these various chatbots to create conversations as part of their coursework. That makes one wonder what a discussion is meant to accomplish and how one may utilise it to spread ideas. It begs the questions of how to evaluate and write a good interaction. Now is a good time for students to reread classical conversations to observe how they function dramatically.
Why not teach students to collaborate with AI writing aides and teach them to think through dialogue if we’re worried about plagiarism? They might learn how to use chatbots to gather inspiration, come up with several angles on a subject, conduct research, and edit the information they find into a logical whole.
At the same time, we must also instruct our students to exercise caution and critical thinking while interacting with AIs and determining the veracity of their claims.
We may all rediscover the rich history and possibilities of this method of interaction through thinking via dialogue.