|
By Dr. Paul Jurcys (Co-Founder, Prifina) Happy Valentine's Day! On this day when the world celebrates love and connection, I would like to share a few thoughts about a rather new paradigm - our relationships with Personal AI assistants and “AI buddies” that are gradually becoming our daily digital companions at work as well as our private lives. I have no doubt that you have been increasingly relying on AI tools such as ChatGPT, or myriads of other AI assistants that helps solve various tasks, I wonder:
 Generalists vs. Personal AI Assistants
The emergence of Siri, Alexa, and ChatGPT reveals a fascinating landscape of technological advancements tailored to our personal, professional, and intimate lives. At the outset, I believe we should distinguish between “generalist AI assistants” and “Personal AI buddies.” Generalist AI assistants, such as ChatGPT and Alexa, designed to provide a broad range of services without accessing personal data. They excel in offering information, performing tasks like setting alarms, answering general queries, and controlling smart home devices through voice commands. Their strength lies in their versatility and the ability to serve multiple users without customization. On the other end of the spectrum, personal AI assistants take customization to the next level. Such assistant are able to tap into a wealth of your own personal data (e.g., data from wearables like Apple Watch or Oura Ring, your own Google Maps starredlocations, and payment histories from your personal credit card records) to offer highly personalized advice, reminders, and insights. These Personal AI buddies can be playful or very practical. Think of a My Fashion AI advisor that knows the styles I like, the shoes I have, as well as the colors I hate to wear. I can talk to my fashion AI advisor about the outfit I should wear for tonights date with my Valentine. In the wake of shortage of human experts (yes, we will never have enough doctors, coaches, or emotional support friends!), specialized AI assistants are becoming increasingly meaningful. Think of an Personal AI doctor who has access to my hearth rate data and sleep information from my wearables, and is capable to monitor different health metrics, suggest activities based on location and preferences, and predict certain events based on my own biometric data. It is not hard to imagine a personal AI finance advisor that knows my spending patterns and history, and can help manage my finances by tracking spending patterns, and nudging to save more wisely to achieve certain financial goals. Assistants According to Functions Performed Simple Personal Avatars help users represent themselves in virtual environments, such as video conferences or virtual reality platforms. They are designed to mimic the users appearance and mannerisms but do not necessarily require deep personal data access. Digital Twins is a step further in terms of their complexity. Digital twins utilize extensive data to simulate and predict outcomes in various scenarios. For instance, a digital twin could use health data from wearables to forecast potential health issues, enabling preventative care. Conversational Interfaces: AI assistants with conversational interfaces, such as ChatGPT, serve a wide array of problem-solving purposes. They can range from general assistants answering any query to specialized ones like AI nutritionists offering certain healthy diet advice or "intimate buddies" providing emotional support. These AI assistants rely on natural language processing to understand and respond to user inputs meaningfully. AI Agents Performing Tasks: These AI assistants are capable of executing specific tasks on behalf of the user, such as purchasing concert tickets or booking appointments. They might access personal calendars, preferences, and financial information to autonomously carry out tasks, streamlining the user's daily routine and ensuring engagements align with their interests and schedules. The Privacy Paradox in the Age of AI Creating personal AI assistants that genuinely understand and cater to our unique preferences and needs, requires a nuanced approach to data utilization and privacy. Let’s take a very practical illustration: what does it take to build a truly personal travel AI advisor? What specific user-generated data would be needed to build a truly personal travel AI? The answer largely revolves around integrating user’s own data from various sources that could help inform the AI to make truly personalized travel recommendations: Accessing Calendar Data: your personal AI travel advisor could benefit by integrating with your calendar data. This connection could offer insights into your available free time, potential holiday slots, and preferred travel durations. By understanding when you are planning to take time off, the personal travel AI advisor can suggest travel options that align perfectly with your schedule. Leveraging Google Maps Data: Further personalization could be achieved by analyzing your Google Maps data. This could include places you have frequently visited and enjoyed, restaurants that align with your culinary preferences, and even destinations you have marked as "want to visit." Such data could enable the personal AI travel assistant to craft travel suggestions that resonate with your past experiences and future aspirations, making every recommendation feel tailor-made. Incorporating Airbnb Booking History: Additionally, integrating our Airbnb booking history could provide the AI with a deeper understanding of your accommodation preferences and budgetary constraints. By analyzing the types of places you have stayed in the past, the personal AI travel advisor can tailor its accommodation suggestions to match your taste and financial comfort zone, further personalizing the travel planning process. However, the aspiration for such a highly personalized AI travel advisor encounters significant hurdles under current data privacy regulations and the reality of data silos maintained by tech giants. These barriers often prevent the seamless integration and utilization of data across platforms, limiting the potential for truly personalized AI assistants. To overcome these challenges, a radical reimagining of the data ecosystem is necessary, where individuals have sovereignty over their data. In this envisioned future, personal AI assistants operate directly on user’s own data, where data is private-by-default, ensuring personalization does not come at the expense of data privacy. This shift towards this more human-centric and user-controlled data ecosystem not only promises to enhance the personalization of AI services but also represents a critical step towards reconciling the competing interests of data utility and privacy. By placing individuals at the center, we can unlock the full potential of AI in personalizing experiences while safeguarding our right to privacy. Reimagining Consumer Experience in the Age of Personal AI’sIt’s time to recognize that we are rapidly entering an era dominated by personal AI assistants who will be seamlessly integrated into many aspects of our daily lives, including our most intimate spheres. It is also obvious that the way how we interact with apps and digital services is set to undergo profound transformation. Rather than having hundreds of apps, we will have one single portal with personal AI agents that have different specializations, and know us deeply (based on our own data). This transformation beckons us to ponder deeply about the nature of our future interactions with intelligent AI assistants and agents and the implications of such relationships on our emotional well-being and the essence of human connections. How Do We Build Trust with AI Companions? The foundation of any meaningful relationship, whether with humans or AI, hinges on trust. Therefore, understanding and building trust with these synthetic agents becomes paramount. Trust in such Personal AI assistants will not solely be about reliability or accuracy but will also encompass ethical considerations such as privacy, data security, and the AI's decision-making processes. How do we ensure these AI companions act in our best interests, and what mechanisms will be in place to safeguard our autonomy and privacy? Transforming Interactions and the Future of Intimacy The proliferation of AI agents will redefine our interaction with the world and each other. These AI companions could become our confidants, advisors, and caretakers, managing not just our schedules and health but also providing emotional support. The depth of these interactions raises questions about the nature of companionship and the role of AI in fulfilling human needs for connection and empathy. Will these sophisticated AI agents be able to discern and adapt to our emotional states, offering comfort and advice akin to a human friend or therapist? Relying on AI for emotional support and companionship introduces a paradigm shift in the understanding of intimacy. As personal AI assistants become more integrated into our lives, it's crucial to consider how this will shape our perceptions of love, friendship, and emotional support. Will AI companions complement human relationships, offering support where human availability is limited, or might they supplant traditional human connections in some aspects? Implications for Human Connection and Love The advent of emotionally intelligent AI raises profound questions about the future of human connection and love. As we form bonds with AI entities capable of understanding and responding to our emotional needs, it's imperative to reflect on how these relationships will coexist with human-to-human connections. Will the convenience and consistency of AI support erode the value we place on human imperfections and the unpredictable nature of human relationships? Or, conversely, could the presence of AI companions prompt a deeper appreciation for the unique aspects of human connection that AI cannot replicate? The journey into a future with personal AI asisstants invites us to embrace the possibilities and opportunities of improving our daily lives. The vision of such AI-powered future also requires us to critically assess and navigate the challenges that lie ahead. Let us embrace this new reality and work together in shaping the future.
0 Comments
Prifina’s CEO Markus Lampinen spoke to Alex Bond on the podcast about the current trends with personal data and how individualized data-driven innovation could be introduced to the insurance services market. TLDR
 Markus: My name is Markus Lampinen, I'm based in San Francisco where I run a company Prifina. We help make products smart. We do this by embedding sensors into different types of physical products. For example, I'm wearing - not that you can see it on the podcast - a really nice kimono-based silk jacket produced by a Tokyo-based company Ichijiku. This jacket has an embedded sensor that can measure a number of different things around, e.g., how the jacket is doing, surrounding data like humidity, etc. Our role at Prifina is to overlay different types of datasets like the sensor data and correlating them with other types of data, e.g., your heart rate when you wear the jacket. Prifina allows companies to build different types of applications that enable them reimagine the end-user experience for how we interact with not only physical products but data. Our vision is that in the future, have more data than ever - how can we use it to create new experiences? Something that's different from what we've had? Alex: That's great. I think that when people hear about sensor-driven technology in clothing, they think about performance, sportswear. But it's much more than that, isn't it? Markus: That's right. We work with pro athletes that want to improve their performance, e.g., top basketball clubs. For them, it's about understanding performance, recovery, sleep, stress, hormonal activities, and so on. From a player’s point of view, how can players themselves understand what's going on with their body and their environment to essentially get better? But it's not just about performance sports, it's also about stress - how do you manage stress? There is a lot of data that we already have, e.g., our sleep data. We can go all the way to, for example, luxury fashion and thinking about how does the clothing that you wear impact your your mood or happiness? As for all this data, we're kind of sitting on this goldmine: we, individuals, have so much data but we're just not using it. It's a huge opportunity. With user-generated data, it is possible to build new types of experiences. Alex: We're now talking at the MENA Insurance Summit. How does this relate to, for example, insurers? Markus: We work with a lot of insurance companies around the world are looking look at preventative health, mental health and wellness. If you can make your policyholders healthier and happier, then they're going to be better customers for the insurance company, because they have very similar incentives as the insurance company does. Alex: Okay, applications are really wide for the insurance industry. … Markus, you explained that it's not just about kind of tracking the data, but it is about the lifetime value of a product that is enhanced by your own data.  Markus Lampinen: Indeed, the lifetime value is one important aspect of a product. How are you maintaining your car? Are you driving it too fast, and so on and so forth. But then, we get into this very delicate balance as well: the data should benefit you as an individual, not the insurance company alone. At the end of the day, for example, when it comes to different types of collectibles, tracing the entire value chain (where everything came from, who actually owns it, etc.) becomes important, especially for maintaining the value and also secondary transactions. This fits into a broader conversation about what is authentic, especially in the digital environments powered by applied AI and deep fakes. We have Barack Obama thats now trying to sell you some skin care product. How do you know that something is authentic? But our focus is very much on reimagining the whole of the consumer experience using data. There are incredible applications as we move forward. For example, cars generate a lot of data. So do you as an individual. What if you could just enter into the car which already knows that, am I a five star driver today? Or that I am only a three star driver today? If you slept really poorly of if you had a little bit more than one drink…? We're not talking about sharing it anywhere, but using your own data for your own benefit. That’s where there's this interplay between all the data that we have and all the things around us. How can we tie them into a unique importable experience going forward? It's a very, very big theme, of course, but that's kind of something that we look at as an inevitable trend. Alex Bond: I think people are just starting to wake up to be protective of that. I’d like to finish with one more question: what does it mean for insurance? I'm an insurance guy. What about applications for the insurance sector? Markus Lampinen: Sure. Insurance emerged as a very clear secondary market for personal data use-cases. Let's take an example. After today’s summit here in Doha, I'm traveling to Switzerland where we're working with our partner that has a sensor embedded into both the left and the right ski boot as well as the helmet. So you can actually get your triangulate the skier and their position to maybe help them improve. Now, performance athletes may care about the angle of tilt on their 760 turn on the mountain. But most skiers don't. But if you start essentially hitting this type of data on skiers, could you create individualized insurance per skier per day in dynamic conditions? Could you give them incentives to ski safer based on those? These are some things that are currently really hard to price dynamically. But if you started having this type of data, could he actually create almost like reverse marketplaces for these types of products? Could you do this on, let's say, a resort by resort basis? You can also use it as, let's say recommendations for routes on different types of ski resorts. You can also use it for insurance products as well. So that's been very much our foray into thinking about starting off in a very, very clear value prop for the individual, but then realizing that there's this entire ecosystem. Alex Bond: Markus, you've been very generous with your time. Thanks for being a guest. You find the entire conversation on , here it is.
On May 3rd, 2023, the Prifina team with its co-organizers at New York Life, hosted a panel on how AI-driven innovation and advances will impact consumer experiences. The event was sold out and we thank all attendees who braved the spontaneous rain in San Francisco to join in person. We’re grateful for such lively audience participation and brilliant questions for our panel, which included Mickey McManus, BCG, Malavica Sridhar, Founder, Northstarre, Markus Lampinen, Co-Founder and CEO, Prifina and moderator Amber Wang, Women in AI / Cerebral Valley.  The panelists together raised various observations and addressed the trends in the current market:
 With such a boom in generative AI, personal data and private use of personal data become fundamentally important. Not only because of privacy concerns but also because of representation; if apps and services are built utilizing individual models that represent the users, we can truly deliver private AI that helps us achieve what we want in life. Like an audience member asked: which human activities should we trust to AI and which should we want to perform ourselves? Questions like these are at the very heart of the debate, but should it not be our choice as individuals based on our preferences? After all, studies show that outsourcing tasks and chores you do not wish to do increases your overall satisfaction in your life. Thank you all for the wonderful debate and we cannot wait to see you soon to showcase more of what Personal AI can deliver, privately, to each of us. |
About PrifinaWe unlock value from personal data, privately. Archives
August 2024
Categories
All
|
RSS Feed
Prifina |
Product |
Community |
Partners |
Research |
Sign up for updates* Prifina will not share it with any third party.
|
The User-Held Data CompanyPrifina is a venture-backed data platform based in San Francisco dedicated to enabling Personal AI Twins and personal data ownership. Prifina empowers individuals to collect, combine, and leverage their personal information, giving them full control over their digital lives. Consumer brands and developers can create AI agents and apps that deliver personalized customer experiences while ensuring users retain ownership and control over their data.
© 2024 Prifina, Inc.
|
|
|
