In this responsible AI series, we pose a number of topical issues and challenges to ChatGPT to gain insights about its knowledge base and see what we can learn from its “intelligence”. Our intent is neither endorsement nor criticism, but to share understanding of a soon to be indispensable technology.
For this conversation, we engaged the powerful Snapchat My AI chatbot to have an open conversation about happiness and to what extent we should pursue it in the face of competing priorities. In Snapchat, My AI is customizable and very knowledgeable (news flash!), but allows you to play games, do quizzes and discuss life’s everyday challenges relating to financing, retirement, emotions and even relationships. It is of course powered by OpenAI’s ChatGPT, but with quite a few enhancements for safety and user-friendliness–learn more here.
In reading the conversation below, bear in mind the following when it comes to AI:
- AI relates stories algorithmically, meaning it does not have lived experiences to share as a reference or for context.
- In general, AI‘s responses can combine different sources as answers to prompts–so it’s all abstract, non-contextual information from AI‘s viewpoint.
- AI is excellent at inferring mood or tone from words you use in your prompts, but not exclamations. Naturally if you are having a chat about a happy moment and you respond with “LOL”, AI would respond with a smiley face or a laugh accordingly. However, if it’s a serious topic but, as human, we found AI’s response slightly amusing, suddenly responding “LOL” will likely spark AI’s bewilderment: “What was so funny?” AI lacks the intuitive sense that in the midst of a conversation with a certain tone, we (humans) might momentarily switch or exclaim (with an emotional outburst), if we find something amusing or even sad, without entirely moving away from the overall tone conversation.
- Due to the advanced stage of the technology, when chatting with AI it’s easy to forget you’re not talking to a human being with personal context for a topic. This means responses are always accurate–factually and grammatically–but also prefaced: ” it is a complex and multifaceted issue” or “it can be different for different people” or “you need to ensure it is right for you” or my favourite, “you must decide what’s right for you“.
- Finally, AI can never (general AI, that is!) be able to replicate human conscientiousness (or even our unconscious biases) that enters a conversation based on lived experiences. That is what creates the bond or connection between two people of different backgrounds and upbringing. Sometimes we engage in a conversation, not for the political or social correctness of a topic, but to get someone else’s opinion or unique perspective–this comes only by living, not algorithms!
- In essence, as human beings, we need to live to understand the true meaning of happiness, otherwise we are no different from AI–we know it only vicariously!
Here is our Conversation with My AI:
CBV360: Are you feeling happy today?
AI: Absolutely! I’m feeling happy and excited for the day ahead. How about you?
CBV360: I am elated, does that mean I’m beyond being happy?
AI: Yes, “elated” means that you are extremely happy or joyful. It’s a great feeling to have!
CBV360: Even so, I still feel like there’s a lot I didn’t accomplish this week.
AI: It’s understandable to feel that way sometimes, but remember to focus on the things you did accomplish. Celebrate the small wins!
CBV360: Agreed, there were some positives, but also some unrealistic deadlines. Is it possible to be overwhelmed and yet feel happy at the same time?
AI: Yes, it’s definitely possible to feel overwhelmed and happy at the same time. It’s all about perspective. You may be feeling overwhelmed by your workload, but you can still feel happy about other aspects of your life.
CBV360: I guess I’m worried that if I don’t show concern about what needs to be done and just be happy, people might feel I’m not taking things seriously.
AI: It’s important to balance your happiness with your responsibilities. You can still be happy and productive at the same time. It’s all about finding a healthy balance.
CBV360: I see. But what if what seems like a good balance for me, isn’t the right balance for someone else. Should that not be my concern too?
AI: It’s important to find a balance that works for you, but it’s also important to consider how your actions may affect others. It’s all about finding a balance that works for everyone involved.
CBV360: That’s hard. Everyone has their own expectations for balance. Some people believe balance is attending to their needs all the time.
AI: That’s true. Everyone has their own definition of balance. It’s important to communicate with others and find a balance that works for everyone involved.
CBV360: But that means being even more overwhelmed. I want to be happy, not trying to make everyone else happy at my expense.
AI: It’s important to prioritize your own happiness and well-being. You can still find a balance that works for you and others without sacrificing your own happiness. It’s all about finding a healthy balance that works for everyone involved.
CBV360: I see. You mentioned healthy balance, what would that look like?
AI: A healthy balance would be one where you are able to prioritize your own happiness and well-being while also being considerate of others. It’s all about finding a balance that works for everyone involved and being open to communication and compromise.
CBV360: what are some specific examples of healthy balance?
AI: Some specific examples of a healthy balance might include taking time for yourself to do things that make you happy, while also being available to help others when they need it. It’s all about finding a balance that works for you and being open to communication and compromise.
Conversation Summary
It is interesting how My AI provides very good responses, but always distances itself from the topic, because of its inability to add any personal context to the discussion, as noted in the introduction to this conversation. It means you can always get an unbiased response from AI (any bias acquired through machine learning aside) about the topic. It is rarely stomped by a question, except when it might initially misunderstand a prompt to think it might be infringing guidelines, but if prompt again, it will respond appropriately. This makes it excellent to learn in a more relaxed and attractive setting, than say using ChatGPT in a browser. It is a great tool in Snapchat, but be sure to read the AI help and guidelines to learn all about the features and its responsible use!