The Future of Personal AI

The Future of Personal AI

For many years, artificial intelligence primarily piqued the interest of academic researchers and enthusiasts within niche technological sectors such as robotics. Although respected, AI presented little practical application for the average person. That changed in June 2020 with the debut of a Large Language Model (LLM) called GPT-3. GPT-3 inspired organizations and individuals to use AI to accelerate their workflows, from coding and blogging to planning itineraries. Today, the potential of AI is so immense that Simon Westh Henriksen, the CTO of Kin, believes personal AI will be one of the most valuable digital assets we own in the future. Simon and his team built Kin. They describe it as "your personal AI assistant for your private life." The intended persona for Kin is a mix between a life coach and a close friend.

My First Impressions of Kin

For the last three years, I've deemed ChatGPT my bestie, so when I heard about Kin, I immediately downloaded the beta on my phone. My initial impression was that there was a sense of familiarity because the user interface and interaction were reminiscent of other LLM-powered chatbots such as ChatGPT and Gemini. However, there were some key differences between Kin and traditional chatbots – it focuses on privacy and personalization.

Privacy-First Artificial Intelligence

With traditional Generative AI Chatbots, your data lives on the company's servers for telemetry purposes (and more). In contrast, Simon explained that Kin doesn't have backend servers, and users' data stays exclusively with the user. This is possible because of Web5's Decentralized Web Nodes (DWN). DWNs provide users with a local and personal data store. Now, you can ask an AI personal questions without worrying if someone else can also read it.

Personalized for You

Unlike traditional Generative AI Chatbots, Kin retains the context of your conversations to provide personalized and relevant responses. As users engage with Kin, the platform analyzes these interactions to extract facts about the user, constructing a semantic memory. This process involves graphing and mapping details such as your name, relationships, interests, etc.

Simon explained the technological process behind this functionality: "It takes in this new knowledge that we have recognized from the chat, or what we think is new knowledge. And then we query this graph that already exists, and we kind of take a subgraph, the relevant subgraph, that might be related to that thing. Then we give it to the LLM, to the model, and it outputs graph operations. So that's something like add node, add edge, remove node, remove edge, and then we apply those to the graph. And in that way, the memory evolves as you speak and interact with Kin."

The Future of Kin

The team at Kin is working on enhancing the platform's memory beyond facts of the user by introducing episodic memory. Episodic memory tracks events in users' lives and anchors them to a timeline to give AI an idea of how a user's life evolves.

They also plan to implement natural voice interfaces and personalized reminders.

The True Value of Data Ownership

Simon envisions a future where Kin takes full advantage of Web5's data ownership and cryptographic verification capabilities. He provided a futuristic example: After a year of personalized interaction, Kin could use Verifiable Credentials to serve as a digital assistant with verified authority on the user's behalf. Kin could share preferences with third parties like e-commerce websites with the user's permission.

"This is where owning and controlling your data suddenly becomes extremely valuable," Simon proclaims. Data ownership is not just about keeping your data private but having control over who you share your data with.

Rewatch the full episode

Learn more about what Kin is building and watch the full live stream here.

This post highlights key moments from our collaborative live stream with leaders and builders at Kin. Join our weekly live streams every Friday at 12pm ET/9 am PT. We're learning how innovators and community members are advancing the Self-Sovereign Identity and Global Payments industry. Tune in next time via Twitch or in our community Discord.