Ivan Crewkov is the CEO & Co-Founding father of Buddy AI, the world’s first conversational AI tutor for youths, on a mission to make sure all students are capable of afford 1:1 English tutoring. After moving to the US from Siberia, Ivan witnessed his preschool-aged daughter struggle to learn English. This inspired him to construct Buddy, a fictional character that children can actually converse with through the facility of generative AI.

Since its launch in 2020, the Buddy app has won several awards and topped the charts within the App Store’s Kids and Education category with over 36M downloads worldwide.

In 2014, you launched Cubic.ai, considered one of the primary smart speakers and voice-assistant apps for smart homes. What were a few of your key takeaways from this experience?

I’m unsure I can take the credit for launching Cubic.ai. I joined the corporate a yr after its foundation and received my co-founder title for my contribution.

Listed below are the important thing takeaways:

• Hardware is difficult, but someone has to do it anyway. Securing enterprise funding for hardware startups is amazingly hard. The one thing that makes things a bit easier is crowdfunding.
• The space of Voice-first products is vast and diverse. What applies to smart homes doesn’t apply to early learning, from technologies to UX design.

Could you share the genesis story of Buddy and the way it originated from your loved ones moving to the USA from Siberia?

With Cubic.ai, I moved from Siberia to the U.S. in 2014 and brought my family with me. My older daughter Sofia began learning English as a second language when she went to a preschool in Mountain View, California, on the age of 4. Sofia struggled to start speaking English for the primary 3 – 5 months in preschool. We were anxious because she couldn’t find friends and play with most of her peers due to the language. We began on the lookout for ways to assist her learn to talk.

It became clear that language apps for youths don’t teach to talk (and every part has stayed the identical over time), and language apps for grown-ups like Duolingo don’t work for youngsters due to the UX. So, we began taking lessons on platforms that connect children with live teachers via video conferencing. Examples are Cambly, VipKid, Novakid, GoStudent, etc. As I observed Sofia learn with live tutors virtually, I saw the good thing about 1:1 attention and energetic speaking practice, but additionally saw the shortcomings of those programs basically.

For instance, as they scale, most of the Online Tutoring Platforms and Online Schools must hire people without pedagogical backgrounds, skills in teaching children, or perhaps a proper English proficiency level. So, to make sure a certain quality of education, online platforms and schools strictly script curriculum and lesson plans, and teachers must use pre-canned exercises, including audio and video fragments. So, unfortunately, on many platforms, tutors principally work like bots.

Still, online tutoring has been the one way for most individuals to learn to SPEAK English, especially in non-English speaking countries. But partly due to the teacher shortage, it’s way too expensive for many families. Learning with live teachers is a premium education service few families can afford.

My co-founder and I got here to the conclusion that AI tutoring is the one scalable option to provide 1:1 English-speaking tutoring to each child worldwide. Soon, we learned that additionally it is one of the best from an academic standpoint. Once we were considering Buddy’s earliest prototypes, we got inspired by research in the sector of Virtual Humans in Education.

Academic studies show animated pedagogical agents’ educational benefits and superiority in comparison with more traditional learning tools and environments. For instance, see Face-to-Face Interaction with Pedagogical Agents, Twenty Years Later, a 2016 article that overviews the sector and cites numerous the relevant material. Here is one quote:

This strengthened our confidence within the multimodal AI tutoring approach. We decided that Buddy can be a multimodal AI tutor – an animated pedagogical agent able to voice recognition and natural language processing. At its core, an AI Tutoring system consists of three essential technologies:

1. Automatic speech recognition (ASR) and evaluation allow us to process and analyze the coed’s speech.
2. Natural language processing (NLP), natural language understanding and dialogue management that processes the content of the coed’s speech and produces the subsequent response. The response consists of each verbal and non-verbal components.
3. Embodied animated virtual character that gives each listening feedback and plays back the system’s response. The character is animated procedurally – the system creates animations on the fly from the NLP response.

All three components are crucial to our approach because only together do they permit us to construct a fascinating, interactive tutor and deliver a successful educational experience.

My daughter Sofia and my co-founder’s son Arseny became Buddy’s first users. Sofia used the earliest versions of Buddy through the first grade.

Several years later, my younger daughter Alisa began using Buddy at three years old when she went to preschool. Now, she is in Transitional Kindergarten and plays with Buddy almost every single day. When Alisa began learning with Buddy, she had several speech issues, so Buddy didn’t understand her more often than not. But after a few weeks of practice, not only her English but her speech improved, as she tried her best to make Buddy understand her.

Why are the legacy ways of teaching a second language so ineffective?

Today, we’re focused on solving particular education problems connected to speech. You may’t learn to talk without speaking practice:

• Most traditional educational tools concentrate on teaching other language skills like reading or writing.
• Language Apps for youths don’t teach speaking skills.
• Some Language Apps for adults today provide speaking practice using AI, but these services don’t work for youths due to UX, safety concerns, and privacy regulations.
• Live tutors are too expensive for many families. Unfortunately, many tutors do not have pedagogical training or aren’t proficient in English.

Buddy is a multimodal AI tutor.

• It’s superior to traditional learning apps because it really works like a live teacher in some ways. Let me quote considered one of our advisors, Dr. Alex Desatnik, PhD, University College London:

• Buddy has some benefits even over human teachers. Buddy doesn’t judge, and for some children, it makes it easier to begin talking to Buddy than to a teacher. That is why today, many tutors use Buddy as an icebreaker that helps children overcome their fear and discomfort and begin speaking the language.

Buddy works to assist teachers, not to exchange them.

I feel it’s very vital to notice this. Buddy may help teachers automate the mundane a part of their job – providing regular practice. We wish to present power to highschool teachers. Buddy is sort of a team of tutors and teacher assistants, working individually with every child in the category and reporting to the category teacher.

Are you able to discuss how Buddy uses elements of gamification to maintain children enthusiastic about learning?

Fun fact: Buddy’s mobile App was downloaded 22 million times in 2023, and over 70% of those downloads were made by children. For kids, our App is a game where they play with Buddy, their talking virtual friend and a preferred Youtuber. Children download the App and persuade parents to pay for a subscription, explaining that Buddy is a teacher.

To make this approach work, we’re designing Buddy as a game with a story and a universe. We work with Hollywood character designers and writers to create Buddy and his story. Now we have a really strong game design team working directly with our educators and turning curriculum and exercises into mini-games in Buddy’s world.

What are another core functionalities that make Buddy so powerful in teaching a second language?

Our core functionality is absolutely focused on Buddy as a multimodal AI tutor:

• Speech recognition
• Conversational AI
• Avatar visual behavior

What are a few of the machine learning algorithms which are used at Buddy?

We’re developing the entire stack of technologies, working together to enable our multimodal AI tutoring approach.

• BSR (Buddy’s Speech Recognition) is a proprietary speech recognition engine specifically to work with accented kid’s speech and comply with regulations like COPPA.
• BLM (Buddy’s language model) — Conversational AI Engine for Children. Protected, fast, and free to operate. It focuses on specific educational functionality and is way less versatile than large language models.
• BABE (Buddy’s Avatar Behavior Engine). This technology generates our character’s visual behavior based on the context of the conversation. Buddy understands when he must smile, change color, or placed on a silly hat.

Many voice recognition systems struggle with accents especially for young children, how does Buddy overcome these challenges?

By developing BSR, our proprietary Speech Recognition technology.

Our unique audience and market required the event of proprietary technology. Buddy must recognize the highly accented speech of young English as a Foreign Language (EFL) learners. One other complicating factor is that beginner students start by learning separate, often short words, that are very difficult to acknowledge without context. Finally, the kids’s market is very regulated, and voice recognition is subject to the Children Online Privacy Protection Act (COPPA) since voice recordings are considered Personal Identifiable Information (PII).

BSR handles kid’s speech with different accents, produced on quite a lot of mobile devices with microphones of varied acoustic qualities and in real-life environments with many sorts of background noise. And it’s COPPA compliant by design.

Working globally, we managed to build up a singular data set to coach our model on. Today, BSR outperforms business off-the-shelf solutions in recognizing and understanding accented kid’s speech.

How do you propose on expanding market penetration to focus on parents who could also be unfamiliar with AI technology?

Buddy began seeing success before AI became a buzzword, and most of our users aren’t the standard early tech adopters. We’re successfully solving a crucial educational problem, and it just so happens that we’re using AI for it.

Still, considered one of the challenges we face is making parents treat learning with Buddy as seriously as with a live tutor — don’t skip lessons, persist with a schedule, etc. The present AI revolution appears to be helping with that.

I’d say that the subsequent big step for us is to begin working more closely with teachers and schools. We’re running a pilot partnership with a faculty in Brazil and discussing partnerships with a dozen more educational institutions.

What’s your vision for the longer term of AI tutors and education basically?

AI tutors are one of the best and the one scalable option to solve humanity’s #1 educational problem – the worldwide teacher shortage. We’d like about 69 million latest teachers to deal with just basic learning needs. For subjects that require 1:1 tutoring, like language learning, the issue is way worse.

The AI revolution accelerated the event of AI tutors, though primarily within the adult segment using off-the-shelf solutions, while early learning stays dramatically underserved. We’re proud to be pioneers of AI tutoring for young children.

Regarding our future, Buddy began as a language learning tutor, but in the long run, it’s going to grow to be an AI tutoring platform teaching a wide selection of subjects to children under 12. Now we have already began rolling out an early version of our first non-language course – the School Preparation Curriculum for U.S. children. We see Buddy because the child’s learning assistant, growing up with a toddler from 3 to 4 years old and teaching multiple courses over a few years.