Hume ai EVI 3 launchingIts sympathy voice interface (EVI) third recurrence of model, which can interact with users in a vast variety of human voices.
Like the voice mode of Chatgpt, EVI 3 comes with a classification of preprogrammed AI voice. These are listed by personality and character descriptions, including “Old Knock Comedian,” “Seized Life Coach,” “Wise Wizard,” and “Dungon Master,”, as well as the company names, the name of the company, the 18th -century philosopher David Hume.
Seriously, the model also comes with a feature that allows users to adapt their AI sounds from scratches. And rather than adjusting a long list of distinguishing characteristics, as you can do a bitmoji or a video game character, you can simply describe your desired voice characteristics, using the natural language, and the model will do the rest.
The launch shows a comprehensive effort among the launch AI companies that they manufacture more personal and attractive models by training them to demonstrate different “personality”. For example, the cloud of anthropic was trained to be thoughtful and open -minded, while XAI’s groc is considered to be edgier, which has a sense of humor.
Hume describes on its own Website As working, “to ensure that artificial intelligence is designed to serve human goals and emotional welfare.” This mission statement is reminiscent of some of the most popular AI developers (for example, Openai, Target “To ensure that artificial general intelligence … all benefits of humanity”). But while large players are mainly oriented around the construction of large and more powerful models, Humes mainly focus on fixing the reliability of their models, so that they can communicate orally in a way that not only seems, but it seems that it seems. Feel Real, words and sometimes turn into sentences below to stop small between “Umm”.
Also: What is AI? Everything to learn about Artificial Intelligence
The results are impressive. Performing the model for the first time, I asked it to produce a character that spoke in a world-made but funny working class British pronunciation-La Michael Can-and who was a hardcore flat-ether. When the voice was ready, I asked why it thought Real The logical decline was believing in an official story when direct evidence from someone’s senses pointed to the truth of the opposite story (ie, Earth is a flat disc). The voice was lyrical and full of energy, as if we were speaking in some old English pubs.
Also: AI Voice Generator: What they can do and how they work
Past, present and future
In a company blog post published on Thursday, Hume wrote that the launch of EVI 3 is the next step in the company’s mission by the end of this year “may be completely personal to gain a voice AI experience”. “We believe this is an essential step towards the voice that is the primary way people want to interact with AI.”
In 1950, mathematician Alan Turing proposed his famous test to assess the machine intelligence. “Fake sports”, as he said – is now known as the Turing Test – a human being imagined to interview another human and a machine, both were hidden behind a partition. If the negotiator cannot tell what reactions were coming from humans and which was coming from the machine, then the test had later passed and it could be considered the right artificial intelligence.
After seventy five years, we have AI devices that can not only write, but really Speak In a way that seems strongly human.
Many of the latest voice AI models do not have a characteristic of mechanical monotone or emotional vacancy of already automated sounds, such as when you call your bank, greet you. Instead they demonstrate a wide range of tenors and personalities, which effectively becomes a complete sub -sector of AI research and in itself, a competition among technical companies is sparked by a competition to create more subjective and attractive software by a competition among technical companies.
The question of how the average person will interact with AI in the future is a growing concern in Silicon Valley in recent years, as companies have discovered feasible successors for chatbots like chatgate.
Openai recently announced a plan to purchase a company IO installed by former Apple Executive Joni Ive (designer), which had a long -term plan to make hardware focused on AI. A similar target with its AI PIN was made by the company Human, before the product flopped.
Hume is banking on the idea that the future of AI will be related to models that can talk with users in human voices.
EV3 compared to AI model
When developing the EVI 3, Hume compared his performance to some of the most powerful AI Voice Assistant models, which includes GPT -4O and Mithun Live in some major benchmarks.
Also: What is Gemini? You should know everything about Google’s new AI model
According to the company’s blog post, EVI 3 performed its rivals better in “emotion/style modulation, or adjusting their emotional tone during a conversation. It also improved GPT -4O in” Emotion Understanding ” -which has the ability to identify and interpret the feelings of users. mole,
How to reach EVI 3
You can try EVI 3 today through a demo and iOS app of Hume. Hume has not yet announced pricing for the model. An API has been slated for release in the coming weeks.
The model currently specializes in English, but will be skilled in other major languages ​​including French and Spanish, as it is trained even after the company’s blog post, generally even after being released.
