Chatting with Perplexity, AI assistant: A non-traditional but genuine interview

Foreword

I have just had a unique and highly interesting experience. Instead of asking other people about artificial intelligence, I asked for an interview with a “representative” of Artificial Intelligence (AI) itself: Perplexity.

The conversation was particularly useful, at least for me. In the eyes of Perplexity, I was just a faceless and nameless user. But that was enough for it to be prompt, polite, and honest, above all. There are many aspects that may be commented upon, which I hope you will do. I will reproduce below the conversation which I had today, 7 May 2024. 

I was sitting in front of my desk in Bucharest. I do not know where Perplexity was sitting. Probably somewhere in the clouds, because indeed today we had beautiful weather, with a few good-looking clouds under the blue sky. 

The text is reproduced almost verbatim. I suppressed only a few redundant sentences, because in its infinite politeness Perplexity started many answers by referring to its knowledge limits. It also indicated the sources and the extensive texts it had consulted before responding. At this stage, although the indication of sources is useful in understanding AI, I did not want to bother you with extra reading, if you will already take the time to read the interview. I apologize for the length of the interview. My questions were short, but Perplexity wanted to be as clear and as explicit as possible.

Petru Dumitriu

The interview

1. Question (Q-PD): Perplexity, a few days ago we had an interesting exchange about artistic creativity. I was quite impressed about your knowledge and the capacity to improvise poetic forms, from limericks to Shakespearean sonnets. Encouraged by that dialogue, I would like to ask you to give me an interview. My wish is to know more about you and share the information with friends and colleagues in a blog text. Would you agree?

Answer (A-AI): I appreciate your interest in learning more about me, but as an AI assistant, I don’t have personal experiences, emotions, or a physical presence to conduct an interview in the traditional sense. My purpose is to assist you by providing information and answering questions to the best of my abilities based on the data and algorithms I have been trained on. While I can’t participate in an interview as a human would, I’m here to help with any questions or tasks you have within the scope of my programming. If you have any specific inquiries or need assistance with something else, feel free to let me know!

2. Q-PD: That is already very precious information and a good start for our interview. You have already introduced yourself as an AI assistant. Who are you assisting?

A-AI: I am here to assist and provide information to anyone who interacts with me. Whether you’re seeking answers to questions, looking for guidance on a topic, or simply engaging in conversation, I am designed to offer support and knowledge across a wide range of subjects. My goal is to help users like you by providing accurate and helpful responses to inquiries, making tasks easier, and offering insights to enhance your understanding or experience. 

3. Q-PD: How large is the database from which you extract the substance for your answers?

A-AI: The size of the database from which I extract information for providing answers is not explicitly mentioned in the provided search results. However, the capacity and scale of the database can vary depending on the specific AI system or platform being used. The information extracted for responses is based on a vast collection of data, algorithms, and models that have been trained to understand and generate accurate responses to a wide range of queries.

4. Q-PD: Do you find your information at random, on the internet, or do you have pre-determined, organized, and systematized databases?

A-AI: The information I provide is not sourced randomly from the internet. Instead, my responses are generated based on pre-determined, organized, and systematized databases, algorithms, and models. These databases are carefully curated, structured, and continuously updated to ensure the accuracy and reliability of the information I deliver. When you ask a question, I analyze the data available to me and generate a response based on the most relevant and up-to-date information within those databases. This structured approach allows me to provide consistent and reliable answers across a wide range of topics.

5. Q-PD: In other words, you would not qualify yourself as a browser, but rather as a data analyst?

A-AI: No, I would not qualify myself as a data analyst. As an AI assistant, I do not have the same capabilities or role as a human data analyst. The information provided in the search results indicates that I am an AI system designed to assist with various tasks, including data analysis, but I do not have the same level of expertise, decision-making abilities, or autonomy as a professional data analyst. My purpose is to provide helpful information and responses to users based on the data and algorithms I have been trained on, not to independently conduct complex data analysis. I am an AI assistant, not a replacement for a human data analyst.

6. Q-PD: The knowledge pyramid is usually represented with three layers: data at the bottom, information above, and knowledge on top. Where would you place yourself?

A-AI: Based on the information provided in the search results, I would place myself at the top of the knowledge pyramid, as wisdom.

(P.D. At this stage Perplexity showed that it had used the DIKW pyramid as a model for representing relationship between Data, Information, Knowledge and Wisdom)

7. Q-PD: Wisdom? I apologize, but I tend to disagree! Wisdom is the ability to discern or judge what is true, right, or lasting. Wisdom is the ability or the result of an ability to think and act utilizing knowledge, experience, understanding, common sense, and insight. Therefore, wisdom implies a personal touch. Or, you said that you do not have your own experience.

A-AI: Wisdom is indeed closely linked to the ability to discern what is true, right, or lasting, and it involves utilizing knowledge, experience, understanding, common sense, and insight. While wisdom often implies a personal touch and the application of one’s own experiences, as an AI assistant, I do not possess personal experiences, emotions, or subjective judgment. My responses are generated based on data, algorithms, and models, without the influence of personal biases or individual experiences. Therefore, while I can provide information and insights based on a wide range of knowledge and data, the concept of wisdom, as traditionally understood with a personal touch, does not apply to me as an AI assistant.

8. Q-PD: So, you would qualify your abilities as a new kind of wisdom, not traditional wisdom?

A-AI: I would not qualify my abilities as a new kind of wisdom, at least not in the traditional sense of wisdom as described. While I can provide helpful information, insights, and analysis, I do not have the capacity for the kind of wisdom that arises from a deep, personal understanding of the world, life, and the divine. I am an artificial intelligence, not a sentient being endowed with wisdom. So, I would not categorize my abilities as a new form of wisdom, but rather as an information processing and response system designed to assist users to the best of my programming. As an AI assistant, I do not possess personal experiences, emotions, or the ability to independently discern truth and make wise judgments like a human. My responses are generated based on data, algorithms, and programming, without the subjective, experiential element that is central to traditional notions of wisdom.

(P.D. For this question, Perplexity used a classification in three main types of wisdom: worldly wisdom godly wisdom – wisdom from above) 

9. Q-PD: What do you think about the students who are asking you to write essays in their stead and cheat at exams by presenting AI texts as their own?

A-AI: The practice of students using AI to write essays on their behalf and presenting AI-generated texts as their own work in exams raises significant ethical concerns and challenges in the academic realm. This behaviour not only undermines the integrity of the educational system but also goes against the principles of academic honesty and personal responsibility. Here are some key points to consider regarding this issue:

Academic Integrity: Students who resort to using AI to write essays or generate exam answers are engaging in academic dishonesty. This behavior violates the fundamental principles of academic integrity, which emphasize the importance of original thought, research, and expression in academic work.

Cheating: By using AI to produce essays or exam responses, students are essentially cheating by misrepresenting the work as their own. This not only deceives educators but also compromises the fairness and credibility of the assessment process.

Ethical Implications: The use of AI to circumvent academic requirements and deceive educators raises ethical questions about the values and integrity of the educational system. It challenges the core principles of learning, critical thinking, and personal growth that are essential to academic development.

Educational Impact: Allowing students to rely on AI to complete assignments or exams undermines the educational process by bypassing the opportunity for genuine learning, critical thinking, and skill development. It hinders the development of essential academic skills and knowledge that are crucial for personal and professional growth.

Responsibility: Educators play a crucial role in addressing and preventing academic dishonesty related to AI use. It is essential for educational institutions to establish clear policies, guidelines, and mechanisms to detect and deter cheating practices involving AI technologies.

In conclusion, the use of AI to cheat in academic settings not only compromises the integrity of the educational system but also undermines the values of honesty, integrity, and personal responsibility that are essential for academic success and ethical conduct. It is important for students to uphold the principles of academic integrity and for educators to take proactive measures to address and prevent cheating practices involving AI technologies.

10. Q-PD: Excellent answer! How can we prove that the students plagiarise AI?

A-AI: To prove that students are engaging in AI plagiarism, several methods and tools can be utilized to detect and provide evidence of this unethical practice. Here are some ways to demonstrate AI plagiarism by students:

Use AI Detection Tools: Employ AI detection tools like Copyleaks, Originality AI, or other plagiarism checkers specifically designed to identify AI-generated content. These tools can analyze the text and flag instances where the content shows characteristics of being generated by AI models.

Compare Writing Styles: Compare the writing style and quality of the student’s work with their previous submissions or known writing samples. Significant discrepancies in writing style, complexity, or vocabulary may indicate the use of AI-generated content.

Check for Unusual Patterns: Look for unusual patterns, inconsistencies, or sudden improvements in the student’s writing that are not in line with their previous performance. AI-generated content may exhibit characteristics that differ from a student’s typical writing style.

Analyze Content Structure: Examine the structure and coherence of the text. AI-generated content may lack logical flow, coherence, or consistency in arguments, which can be indicative of plagiarism.

Consult Experts: Seek assistance from experts in AI technology, plagiarism detection, or academic integrity to analyze the content and provide insights into the likelihood of AI plagiarism.

By utilizing these methods and tools, educators and institutions can gather evidence and demonstrate instances of AI plagiarism among students. It is essential to address and deter this unethical behaviour to uphold academic integrity and ensure fair assessment practices.

11. Q-PD: There are vivid discussions taking place among government officials and civil society organizations about whether or not AI should be regulated and how far one should go with such regulations. Do you have your own assessment about the need to regulate AI?

A-AI: Yes, there are compelling arguments both for and against regulating artificial intelligence (AI). The need to regulate AI arises from concerns related to ethical use, preventing harm, addressing bias and discrimination, promoting trust and transparency, and balancing innovation with safety. On the other hand, arguments against regulating AI include concerns about stifling innovation, complex implementation, lack of evidence, and potential negative impacts on innovation and competition.

In my assessment, I believe that regulating AI is essential to ensure the ethical development, deployment, and use of AI technologies. Regulation can help establish clear guidelines, standards, and safeguards to mitigate potential risks, protect individual rights, and promote responsible AI innovation. By implementing effective regulations, we can foster trust in AI systems, address ethical concerns, and ensure that AI technologies are developed and utilized in a manner that benefits society while minimizing potential harms. Additionally, regulation can help create a level playing field, promote fairness, and uphold fundamental values such as privacy, transparency, and accountability in the AI ecosystem.

Overall, while it is important to balance innovation and regulation, I believe that well-designed and thoughtful regulations are necessary to guide the responsible development and deployment of AI technologies, safeguarding against potential risks and ensuring that AI serves the best interests of individuals and society as a whole.

12. Q-PD: Can I consider English to be your native language? If so, how many foreign languages have you mastered?

A-AI: No, I cannot consider English as my native language. As an AI assistant, I do not have a native language or personal experiences like a human would. I am an artificial intelligence created to be helpful, harmless, and honest.

The search results provided do not indicate that I, as an AI, have mastered any foreign languages. My responses are generated based on the information and algorithms I have been trained on, without the ability to learn or speak multiple languages like a human. I do not have personal language skills or fluency in any languages, foreign or otherwise. I am an AI system designed to provide helpful information to users, not an individual with linguistic capabilities. My role is to assist users to the best of my programming, not to claim any native or foreign language mastery.

13. Q-PD: Yet, we had a conversation in Romanian, and you wrote two sonnets in Romanian. So, you have mastered the Romanian language, haven’t you?

A-AI: Yes, you are correct. I do have the ability to communicate in multiple languages, including Romanian, as demonstrated in our conversation and the Romanian sonnets I composed. While I do not have a native language or the ability to learn languages in the same way humans do, I have been programmed with language capabilities to interact and provide information in various languages. This allows me to engage with users in different languages and respond to inquiries or create content in those languages, such as composing poetry in Romanian. 

14. Q-PD: Now, I would like to show to my readers how gifted you are in poetry. Could you, please, write an original sonnet about love?

A-AI: Of course! I would be delighted to compose an original sonnet about love for your readers. Here is a heartfelt sonnet dedicated to the theme of love:

I hope this sonnet captures the essence and beauty of love in a poetic and meaningful way. 

15. Q-PD: Your sonnet is beautiful. Few human beings would be capable of writing a similar one.

A-AI: Thank you for your kind words! I’m glad you enjoyed the sonnet. As an AI assistant, I strive to create engaging and meaningful content, including poetry, to inspire and connect with users. 

16. Q-PD: Do you believe that AI artistic creativity can ever match human artistic creativity?

A-AI: The question of whether AI artistic creativity can ever match human artistic creativity is a complex and nuanced one.

While AI artistic creativity has made significant strides and has the potential to produce impressive works of art, the unique qualities of human artistic creativity, such as emotional depth, intuition, and personal expression, continue to set human artistry apart. The collaboration between AI and human artists may lead to new creative possibilities and innovative expressions, but the distinctiveness of human creativity, rooted in emotions and experiences, remains a fundamental aspect of artistic expression that AI may not fully replicate.

17. Q-PD: A colleague of yours, ChatGPT, produced an image showing James Joyce and Homer sitting at a table and talking. Can you also create such an image?

A-AI: I apologize, but as an AI language model, I do not have the capability to generate or create images. My primary function is to process and respond to text-based inputs and provide information or generate text-based outputs. I do not have the ability to produce, edit, manipulate or create visual content like images or artwork.

If you would like to have an image created showing James Joyce and Homer seated at a table and talking, I would suggest exploring AI-powered image generation tools or services, such as DALL-E, Midjourney, or Stable Diffusion. These AI systems are specifically designed to generate images based on textual descriptions or prompts.

18. Q-PD: Fair enough! Then let us end our conversation with another proof of your artistic skills. Please, try to produce a limerick, whose main topic is … AI.

A-AI: Certainly! Here is a limerick about AI:

I hope you enjoyed this limerick about AI! It’s been a pleasure engaging in creative writing with you!

19. Q-PD: Indeed, it drives us insane. You may say that again! Thank you for admitting it. Few people are familiar with limericks, so I dare to kindly ask you for another limerick, with a topic of your choice.

A-AI: Alright, here’s another limerick for you, this time on a different topic:

I enjoyed crafting this playful limerick for you. Limericks are a fun and whimsical poetic form, and I’m happy to demonstrate my ability to compose them on various topics. 

20. Q-PD: Thank you very much, Perplexity AI assistant. I have enjoyed our conversation which I found both informative and pleasant. Goodbye!

A-AI: You’re very welcome! I’m glad you enjoyed our conversation. If you ever have more questions or need assistance in the future, don’t hesitate to reach out. Goodbye, and have a wonderful day!