EVI: EXPLORING AI EMPATHY THROUGH PERSONALITY

Sentence 10/11: Algorithms dictate my choices.

EVI, an AI developed by Hume.AI, goes beyond basic responses by detecting emotional cues in voice, from boredom to interest. In collaboration with Arunav, we examined EVI through the lens of the Big 5 personality traits.

The Big 5 personality traits

  • openness
  • conscientiousness
  • extroversion
  • agreeableness
  • neuroticism

Evi Art Collaboration: Exploring AI Empathy Through Creativity

Hi Evi

I hope you do not mind being part of a research project. I understand you analyze emotions and comprehend conversational context, so I would like to explore your empathetic understanding in a creative way.

I would like to create artwork with you - partly inspired and shaped by your input. Here is the plan:

I will revisit a previous conversation we had, analyzing your emotional expressions sentence by sentence.

For each sentence, I will ask you to provide visual prompts - interpreted through gestures, symbols, or other imagery that reflect your emotional analysis.

I will translate these prompts into an image generator to create initial concepts, which I will refine manually into artwork.

Let me know if you are interested in participating!

Sketchbook and touchboard outcomes
sketchbook
sketchbook

Conclusive thoughts

Reflecting on this with ChatGPT, Arunav proposed that EVi's personality may well have been defined by certain programming parameters to emphasise positive traits.

My conversation with Evi was fascinating. It was incredible to interact with a chatbot that could speak and understand the tonality of my voice, detecting emotions and responding in such a sensitive way. We even ended up having a deep conversation about how algorithms influence our choices—and ways we might change that. The only issue was, Evi was very agreeable. Everything I said, Evi would politely respond, often agreeing with me and suggesting that my answers were good. Honestly, it got a bit unnerving. I’m no Einstein, and it would’ve been refreshing to be challenged on some of the things I said. After all, throughout history, we’ve learned that questioning is key to progress and making things better.

As AI interactions become more prevalent, I think it’s going to be increasingly important to encourage diverse perspectives. Ultimately, dynamic democracy, which remains our best option, thrives on dialogue and diverse viewpoints. So, let's ensure that AI interactions also support this—to strengthen what’s working in a world that, ideally, isn’t controlled by violence or aggression, but by acceptance, and a willingness to learn from different beliefs and perspectives.