Personal Guardrails for AI Use— Freedom and Responsibility

In my last post, “AI, Art, and the Human Soul,” I ended with the idea that we should use artificial intelligence (AI), but do so responsibly. What does that look like and what kinds of personal guardrails can we implement to protect ourselves from AI’s downfalls?

I’ve spent a surprisingly large amount of time thinking through this question because I’ve been experimenting with AI for years and I can feel its pull to be lazy. I think that may be the biggest problem with using it. So that leads me to my first personal guardrail:

1. Do the Work That Forms You

This may be the most challenging and important guardrail!

AI can do so much work for us, but to what end? What is AI freeing us to do? Are we using that freedom to become more capable, creative, disciplined, and connected?

If AI is freeing us up to be entertained, it’s probably having a net-negative impact on our lives. However, if AI is freeing us up to do the kind of deep formative work that’s turning us into who we want to become, that’s golden!

Freed from Physical Labor

I think the Industrial Revolution gives us a great analogy here. Prior to the Industrial Revolution, most people had very physically demanding lives (picture yourself walking to town and hand-washing your clothes at your homestead). Sometime between then and now, our lives have become very sedentary; most people work at a desk and our physical health suffers for it. So now, with this freedom to do different kinds of work, comes the responsibility to do special things to take care of my physical health. In other words, I need to exercise! Gyms, as we know them today, didn’t exist prior to the Industrial Revolution because we didn’t need them. When all these labor-saving devices freed us up to do desk work, they also created a situation where we need to take responsibility and do the physical labor our bodies need.

I’m sure you already see the parallel to AI. AI can save us from having to do a lot of mental labor. However, if we’re lazy and let AI do all our thinking for us, we’ll become mentally weak. If we want to retain abilities like remembering, reasoning, writing, and solving problems, we need to practice them intentionally. We have to be responsible and engage in meaningful, challenging mental work.

Focus on Human Formation

So how do I decide which work to do and which work to let AI do?

We still need to do some mental work because it forms us as humans, so I try to choose to do formative work myself and have AI do the less-formative work. This lens of human formation is the best way to maintain the personal guardrail of “don’t be lazy,” so let’s examine it a little closer.

The work we do and the disciplines we practice form us into who we become as humans. (I believe this is why God gave us the Spiritual Disciplines.) Our disciplines and habits form our human character.

For example, the only way to develop patience is in a situation that requires patience. We have to sit in the difficult, often frustrating, situation so that we can become better at dealing with it. If I habitually use AI to escape frustration, confusion, or slow progress, I may also escape the opportunities through which patience and perseverance develop. That struggle helps our character grow.

In the same way, creating and appreciating art forms us. We spend time looking at art and develop our ability to notice details. We ask ourselves what beauty is and develop an appreciation for beauty. We meticulously draw a subject and strengthen our fine motor skills and our patience with details.

So ask yourself questions like:

• Am I trying to learn this skill or merely complete the task?
• Would outsourcing it prevent me from understanding the result?
• Is judgment the valuable part of this work?
• Is this a skill I want to retain?
• What is AI freeing me to do instead?

Human Relationships

Another way AI can make us lazy is if we shift to spending less time in relationships with humans. If you haven’t used AI at all, this may seem like a weird idea, but it turns out that AI is remarkably good at simulating being a very agreeable human. A chatbot will typically be polite, listen patiently, and tell you the things you want to hear. (You can modify this default by adjusting the personal settings in your AI account.) Your AI won’t complain about its own ‘life’ or get hurt by your angry outburst.

Human relationships, on the other hand, are messy. We are selfish creatures and we’d all like the world to revolve around us. A chatbot can simulate the pleasant features of a relationship, but it cannot participate in a reciprocal relationship with needs, vulnerability, responsibility, and sacrifice. However, there’s something about human relationships that is essential to human flourishing. So, while it’s really ok to talk to a chatbot about what’s on your mind, don’t let it crowd out the relationships with real humans.

Some warning signs include consistently choosing AI instead of talking to someone, avoiding necessary difficult conversations, becoming less willing to accommodate other people’s needs, and treating simulated agreement as genuine intimacy.

Don’t skip the formative work of being in relationship with other humans.

2. Don’t Be Dishonest

I understand that our culture is in the middle of a giant shift and there are some rabid anti-AI people out there. It may feel scary to admit you’re using AI, but hiding it will feel worse. Once you have thought carefully about your choices and established responsible guardrails, don’t be evasive or apologetic about using it. Be honest about what you do, how you do it, and why. I would rather find people who share my values, or who are at least willing to make careful distinctions, than conceal my choices to win the approval of people who have already decided that all AI use is wrong.

How can you decide what to disclose? I think it’s important to try to distinguish between ordinary tool use and material assistance. I think it’s a good rule of thumb to disclose your AI use when:

• Someone reasonably believes the work was entirely yours
• AI made a substantial creative or intellectual contribution
• School, workplace, contest, or publication rules require it
• Hiding your AI use would give someone a false impression of your skill, effort, authorship, or process

3. Take Responsibility

AI can appear confident, accurate, and skilled. However, it’s occasionally wrong or willing to do something unethical. So, you need to treat it like an eager, inexperienced intern (or your crazy Aunt Sally who always seems to know everything).

We have developed a strange double standard: a human mistake is ordinary, but an AI mistake becomes a cautionary tale repeated for years. It’s not helpful to try to hold AI to a standard of perfection that we don’t expect from people. Also know that AI’s accuracy varies by task; it’s just better at some things than others. Treat it as a useful but fallible tool, and match the amount of verification to the cost of being wrong. You are responsible for any content you create with AI, just like you’re responsible for the car you’re driving. So check its work! Read over that email it drafted, inspect that image it created, and double-check that statistic it gave you, just like you would if an intern did the work.

You’re also responsible for the moral choices involved in how you use AI because AI doesn’t have the inherent moral compass you do. Don’t have the tech do things you’d be ashamed to do; it’s still you doing it, just with a more powerful tool.

Privacy

One specific area that kinda falls under this heading is privacy. The more information you give AI, the more helpful it becomes, but what if that information wasn’t yours to share? Be responsible with other people’s sensitive information and don’t give it to AI without their permission.

Ask yourself these questions as you try to decide what information to share with AI:

• Is this information mine to share?
• Does AI need the identifying details?
• Could I use a fictional name or generalized summary?
• What harm could occur if this information were retained or exposed?
• Have I checked the privacy settings and policies for this particular AI service?
• Am I giving the tool more access than the task requires?

Freedom and Responsibility

I’ve intentionally called these “personal guardrails” because the government should not try to regulate the balance we must find between freedom and responsibility (I know, that’s another post for another day). You have to decide what your personal guardrails will be and have the discipline to keep to them.

However, children are not yet ready to bear the full responsibility of total freedom. It’s our job as parents to help our children grow in virtue and character by guiding them through these choices. So while we teach these guardrails to our kids, we also police their work. To do this, we must be clear about what they are and are not allowed to use AI for in their work and hold them to that standard. We’re training them in judgment and self-discipline when we do this.

The use of AI comes down to a balance between freedom and responsibility. The more choices we have in how we get stuff done, the more responsibility we have in making those choices.

---

Note: Yes, I choose that em-dash in the title myself. My AI doesn’t even use em-dashes because I’ve personalized it. I’ll tell you how in the next post!

---