Think about the most impactful and life changing decisions you’ve made—choosing a career, deciding whether to move cities for a job, or leaving a relationship. You probably approached these monumental choices with a mix of gut feeling, advice from those closest to you, and perhaps a coin flip for good measure. It’s a bit like before google maps, we set out on a cross-country road trip relying on a faded map and the occasional advice of strangers. This primitive methodology is met with uncertainty, mistakes, and is just simply ripe for disruption.
In an era where technology has streamlined everything from ordering groceries to solving advanced medical problems, one significant area remains oddly untouched by the sweeping advances of AI: personalized advice for life’s pivotal decisions. I am here to argue why the most groundbreaking, impactful, and lucrative application of AI will be in guiding us through small and big life choices and how these will affect you exponentially more than any other type of AI.
The Primitive Present - Guessing and hunches
Our current approach to making life altering decisions is surprisingly rudimentary. Despite living in a digital age, when faced with significant choices, we often revert to the most basic methods: lists of pros and cons, consultations with friends and family, and maybe reddit or quora for the tech savvy. This is akin to navigating a new city without Google Maps; we will look back and wonder how we ever got by with such useless solutions. It is inevitable that in the future, we will be approaching life decisions with some sort of analytics or AI that will recommend useful and data validated advice.
Even a 1% improvement on our life decisions, over time, lead to vastly better life outcomes.
The AI Advisory Gap - Personal AI’s are not yet Personalized
The potential for AI in offering personalized, actionable advice is immense—yet largely untapped. Existing AI, including the likes of ChatGPT and Pi, has made strides in simulating human-like text based on vast amounts of web data. They make great assistants, but terrible advisors. When I ask it how to lose weight, I am met with a answer that you will also receive. This is not how advice works. We all respond differently, thus we all need individualized advice. Even with chatgpt’s recent announcement of memory, these ai’s are still not truly personalized. ChatGPT can only tell us what people generally do in certain situations and pretty much summarize advice from its data trained on quora and reddit. It can't predict the best course of action for an individual based on their unique circumstances, personality, and the likely outcomes of different choices. In fact much of the training data for these AI systems, is faulty advice. It’s unclear if the advice on Quora and reddit actually worked out, so these Personal AIs not only are impersonal, they actually have faulty advice in their core training data.
The trillion dollar dataset - How much would you pay for useful advice?
Present-day AI systems fall short in offering personalized advice, primarily because they lack comprehensive access to three crucial elements: 1. the full context of the individual, 2. the specifics of the situation, and 3. the outcomes of the advice provided. Without all of this, there is not enough information to properly train an effective AI The transformation of AI into a potent advisor is contingent on amassing and leveraging data that encapsulates the complete scope of the situation, the individual involved, and the results of the advice. This dataset might encompass a variety of information, such as demographics, personality types, genetic data, and social media activity. While there are numerous strategies for tackling this challenge, the core solution revolves around acquiring a high-quality, comprehensive dataset that includes outcome data, which is presently non-existent online. With such data at hand, the development of a personalized advice-giving AI becomes feasible. Initial implementations could be as straightforward as a finely-tuned language model that incorporates information about the individual and their situation into every inquiry. Alternatively, more advanced systems could employ retrieval-augmented generation or employ clustering and classification techniques. The sophistication of these AI models is expected to increase over time, but the starting point invariably remains the acquisition of robust data.
The more people that use the AI and add to its dataset, the smarter and more useful it will become.
Conclusion
It’s clear that the most profound impact of AI on our lives may not come from beating humans at chess, making higher quality ai images, or driving cars to work, but from guiding us through our extremely impactful life’s decisions. It does not need to be perfect, just better than the guessing we do now. This AI will consider your personal history and personality type, your current situation, and data from thousands, if not millions, of similar people and decision points to offer advice with unprecedented relevance and accuracy. AI has the potential to guide individuals in achieving fitness goals, overcoming addiction, advancing careers, and finding love. This expansive reach, touching on every facet of human life, underscores AI's potential to act as a transformative tool for personalized guidance, making it the most impactful yet under discussed application of technology in enhancing the quality of human life.
The true value of AI for the average person lies not in its ability to surpass technical benchmarks or solve complex math problems, but in its profound impact on daily life through evidence based, personalized advice and support… Advisors, not Asssitants.
Part 2 can be found here.