Messaging about AI is hyperbolic: “AI will save us!” or, “AI will doom us all.” Neither is accurate, or helpful. This is because AI is a tool. Like any tool, it is good at solving certain types of problems. And, again all tools, it can be misused, or used irresponsibly, in ways that cause harm. Whether or not AI is useful depends on what problems we tackle and how we tackle them. This often gets lost in the hype and the desperate race to “win” AI (despite no one really knowing what that means).

What’s missing in the AI conversation is an honest assessment of where the technology is now (let’s be real, it’s not yet a life-changing technology), and what we need to do to make it useful in solving meaningful problems, particularly for folks outside the tech industry. I work in this industry because I genuinely believe that we can get to a place where AI can help us solve the biggest challenges facing humanity, from climate change mitigation and systems resiliency to management of renewable energy electrical grids to helping us identify cures and medicines for diseases and cancers and other disorders.

What’s real now and what’s hype?

Today, AI is generally good at solving problems in the realm of mundane and time-saving tasks, things like summarizing, organizing, and as a thought-partner. I break this down into 5 areas across personal and work-related tasks:

1. Brainstorming: AI is incredibly good at generating ideas. Not good ideas, per-say, but lots of ideas. This is where the human comes in: we are much better at recognizing what’s a good, or useful, idea, or at least one worth pursuing.
   1. A sub-point to this one: AI is also very helpful at generating synthetic (i.e., fake) data for testing or storytelling purposes.
2. Thought-partner: AI is also good as a “rubber ducky” – that is, as an entity that we can talk to and explore ideas or questions. Sometimes we get unstuck just by having a conversation (even if it’s one-way) because it forces us to put our problem into words. AI lends a helping hand to this process by leveraging its broad swath of knowledge from training and ability to search the web.
3. Summarizing: Taking lots of content and giving us a brief, actionable summary
4. Organizing and formatting: Taking notes or other content and adding structure.
5. Translating and adapting content: Translations between human languages (e.g. Spanish to English) or adapting content for different audiences, including ourselves (e.g. “break this paper down into simpler terms so I can understand” or “create multiple version of my lesson plan for students with varying reading levels”).

Everything above is useful (I really, really dislike formatting), but they don’t disrupt our ways of working or creating.

What is a game-changer is that AI is increasingly proficient at writing code.

AI-generated code is not necessarily “good”* code, but it’s fast. With the right approach and guardrails (e.g. see Spec Driven Development or SDD), you can cajole the AI to follow best practices and patterns to get functional software at a fraction of the time it typically takes – think minutes or hours instead of days or weeks. This is true even for complex software programs like an eShop or scalable database. What’s more: you don’t have to know how to write code to build something real.

*”good” code is a heated topic of debate, although there are certainly best practices. AI-generated code might be overly lengthy, duplicative, messy, etc., but if a human never has to see the code…. Does it matter?

The massive value-unlock with AI: generating code

In the next year, we will start seeing AI tools that can generate functional software. There are a few out there today, mostly in the realm of UI-generation, and what will change is that folks will be able to generate software using natural language (i.e., human language), and we will never have to look at code.

This is a massive value-add for every single person, even for professional developers (or software engineers aka SWEs). The industry will change – is changing – which is why it’s difficult for folks new to SWE to land a job in big tech. The skills people need are less deeply technical and more design and systems thinking. Like all technological revolutions, today’s jobs will change: some will fade away, many will shift, and new ones will be created.

But so what, you say, you don’t work on software, you’re not a SWE or in the tech industry, why does this matter to you? We all use digital systems every single day: a smartphone, a laptop, heck even our cars are largely digital systems. You can use AI to create the software that you want, to build things in support of your personal goals and relationships, to reduce reliance on corporations that constantly tell us to consume, who tell us that greed and money are the only things that matter.

We can use AI to build systems rooted in love and care because now we get to decide what we want. Until today, the software industry has been a largely gated community – relegated to those with the time and resources and specialized knowledge to define the systems that our diverse, global population use every day. AI generated code breaks down those barriers: anyone can use AI to write code and make their own apps.

Don’t like that social media app? Build your own and share it with your friends and family! Frustrated by the ads you see or the data that is taken from you and sold in those lil’ app games you play while waiting for the bus? Build your own game! Annoyed at the community forum that wants to require facial recognition software, despite it being known to be inaccurate and racist? Build your own and let it be known that we now have power to vote with our digital feet because we now have options.

OF course, I’m simplifying things a bit – we’ve talked about AI-generated code but not so much what it takes to take code and deploy it, i.e., make it (publicly or privately) available to other folks. However, we can all learn those skills, especially when we are motivated to solve a problem. We can pool resources to support grassroots efforts like a DIY social media app – building better systems and community along the way.

What do we do to build better systems with AI?

This most important thing we can do to ensure that everyone sees value from this technology is to start using it with intention. It is easier than ever before to bring an idea to life using AI – whether you need help with market research, brainstorming, or software generation, AI is a tool that will help you through the process.

What problem do you, your friends or family, or your community have that you are passionate about? Pick an AI platform and start exploring solutions for how you might go solve that problem.

Note that not every problem benefits from a software solution, and we should be intentional in what and how we build with AI. However, there are many problems we can solve with AI, especially when we start considering how to empower individuals, how to build systems that are rooted and centered on mutual love and support, how to bake-in privacy and accessibility.

This is an opportunity of a lifetime to decide how we want the future to look. I believe that the vast, vast majority of us want a society that values equality, freedom, love, and kindness. We want a society where we support and care for one another, where we solve problems and seek solutions, where we work together for the benefit of all living beings.

Let’s go build it.