Keeping good company with EVI, Your Empathetic Voice Interface

News Whip

A new data poisoning tool is helping artists fight Gen AI. Humane founders are trying to win over skeptics of their new AI pin (I’m a skeptic but also weirdly bought the Rabbit), and an AI researcher is taking on deep fakes during elections.

AI is helping business leaders pioneer new frontiers.

Tune into the Digital Data Design Institute at Harvard Leading with AI conference on May 7 to learn how your organization can capitalize on digital transformation. Leading with AI is chaired by Harvard Business School professors

Karim Lakhani Tsedal Neeley, and Mitchell Weiss and is presented in partnership with the Digital Data Design Institute at Harvard, a pioneering hub dedicated to the study of transformative technologies. Learn more here:https://bit.ly/HBS_conference I’ll be a speaker at this conference.

Meet EVI.

EVI is an empathetic voice interface. It’s an empathic LLM. The creators, Hume, are pitching this as the first emotionally intelligent voice-to-voice AI. It monitors, predicts and matches your mood. Hume is a “research lab and technology company. Our mission is to ensure that artificial intelligence is built to serve human goals and emotional well-being.”

You can have a surreal audio conversation with it and a back-and-forth that places your emotional needs, feelings and attitudes first. The AI reads the sound of your voice and categorizes the mood in real time. It can also perceive your emotions if you open up the facial expression function.

• It’s really cool and even creepy cool, but not quite there for me
• The voice is kind of annoying
• The sensitivity to any sound cuts off the conversation, so it’s not as smooth. Easy to get annoyed, and then it reads and tells you that you’re sounding annoyed…
• Too perky for me!
• It’s only the demo so I’ll wait and see how the $50 million investment goes

Try it and tell me if you think it can replace therapy and mental health services. My response is, not yet. But there is someone to talk to.

We envision a future where AI systems use scientific approaches to fulfill human needs. Emotional intelligence is the missing ingredient needed to build AI systems that proactively find ways to improve your quality of life.

The above is a representation of the vectors where the speech patterns it has trained on reside.

I now speak fluent Arabic, by the way. Overnight.

Other than the coolness of that, this tool means most major languages can be translated using GenAI tools like HeyGen. Lips moving kind of in unison with the words of the language. Many hours, lots of dollars and logistics were saved like this. It took us about 15 minutes to handle the translation part. The Rundown team is looking at this tool for translations to different markets in the work we do. We also recognize we need to run the AI translations through a human to make sure the translation is correct. That’s a layer that requires an editor and a cost.

1. Transparency Breeds Trust: Be open and honest about your use of AI in content creation and distribution. Explanation: By clearly disclosing when and how AI is used, you build trust with your audience and maintain accountability for the content you produce.

2. Verify, Then Amplify: Establish rigorous, multi-step verification processes for all AI-assisted content. Explanation: In an era of AI-generated misinformation, thorough verification—including source validation, metadata analysis, and human review—is essential for maintaining accuracy and credibility.

3. Proactive, Not Reactive: Anticipate and monitor AI-related threats to stay ahead of the curve. Explanation: By investing in threat detection tools, collaborating with experts, and regularly updating your defenses, you can proactively combat AI-generated misinformation and maintain the integrity of your content.