What is the value proposition of AI in your personal life?
Would you, or do you, pay $20 a month for an AI chatbot? That's the question companies developing frontier models (e.g., OpenAI, Anthropic, and Google) hope you'll answer in the affirmative. Even that 20$ figure is likely lower than what they truly hope you would pay to access their tools.
Until then, they either need to raise money from investors or use other monetization strategies (e.g., ads or corporate sales), and even then, they are losing money. Certainly, there is a clear promise of what AI may deliver in the future for the corporate sector (but, at least currently, it hasn’t really improved output). Either way, people recognize the value proposition of AI in the workplace. For example, this could mean having a personal assistant handle many of your tasks, leaving you to focus on more critical ones, or even full automation of positions.
The same can't be said for AI in the personal world. Sure, you could use chatbots to help write a text to a loved one, or maybe in the future have an agentic model buy your groceries (this is possible now, but buyer beware: risks include having your banking information taken or the wrong items purchased). But even then, it takes real imagination to find use cases that don't involve work.
My academic, professional, and personal life is focused on AI. People often ask me what AI can really do. Everyone has heard about its potential, but only in abstract ways. There are people who claim it can automate your life (it can't) or make it easier (maybe, depending on how you define 'easier'), but these are just vague ideas. So naturally, people come and ask me how they can implement it for themselves. In a professional context, I have a variety of potential approaches for them to consider. But when it comes to personal life, I just explain how I use it — and maybe they can get something out of it.
In my case, I pay for several AI model subscriptions because I’m interested in the technology, not because they dramatically improve my daily life. For example, I don’t use AI agents to buy groceries or do my taxes (though some people do). Personally, I use Perplexity and Google to answer questions. I use Gemini built into Google Maps for hands-free navigation. Claude helps me find weekend activities in Ottawa. All the models have assisted me—sometimes successfully—in troubleshooting issues, like fixing my printer. I also use Spotify’s AI DJ occasionally, although it doesn’t always work, so I often use shuffle. That’s the extent of my use. I don’t want an AI model to be my therapist or a chatbot that tells me I’m always right.
Now, let's contrast this with other technologies. For example, the internet and smartphones. Sure, there were people who were hesitant, but inherently, most people understood some of the utility behind these new technologies. The Internet brought social media, and, again, people flocked to these platforms — without really being convinced.
The same can't be said for AI chatbots. Even though the value proposition is unclear, we are still paying dearly for these frontier models. AI models need data centres, which consume significant water, PC RAM, power, and land. Not to mention, people are constantly bombarded with overblown, pessimistic, and optimistic views on these tools — which are almost always what-if scenarios or dramatizations of the current reality — rather than a closer approximation of the real picture. People talk about the internet, phones, and social media — not in this overblown manner, but by highlighting both the pros and cons of these tools.
Don't get me wrong, I find AI incredibly promising and use it in my professional endeavours. But at least for now, I am unsure of the value proposition these tools offer in the personal realm. This may mean that, eventually, frontier-model development companies will shift their priorities away from the personal altogether and focus only on the enterprise. We may already be seeing this. OpenAI says it's going to cut back on ‘side quests’, which may end some of its attempts to delve into the personal.
Do you have personal use cases for these tools?
Take care,
P.S. This post draws from this Vergecast episode. I highly recommend you give it a listen.