Shift The Default

What if every lawyer’s first question about AI wasn’t “is it good enough?” but “who might it help me reach?”

A few weeks ago, an appellate court issued several decisions in termination of parental rights cases that troubled me. In each one, the court – in split decisions – affirmed the permanent destruction of a child’s relationship with their parent. I saw issues worth raising and arguments that deserved a second look, the kind of problems that have, in my experience, led the State Supreme Court to get involved.

So I did what I’ve always done. I reached out to the appellate lawyers handling the cases to see if they’d consider seeking leave to appeal.

The responses came back quickly. They were polite enough, but the message was clear:

“I don’t get paid to do it.”

“I don’t have time.”

And that was that. Parents who had just lost everything would get no further review – not because the issues lacked merit, but because no one was willing to raise them.

I understand the responses. Court-appointed appellate lawyers often get paid a flat fee that doesn’t cover work at the supreme court level. They carry enormous caseloads. Many of them are doing the best they can inside a system that has never resourced this work adequately. We’ve set them up to fail.

But understanding the reasons doesn’t make the result less devastating. A parent has lost their child. Permanently. And the system’s answer is: sorry, we’re out of time and money.

Our Capacity Problem

This is not an isolated problem. It is the defining reality of legal practice for anyone working with those left behind.

In Child Welfare Appellate Clinic at the University of Michigan Law School, we get calls from families in crisis looking to appeal problematic decisions; parents whose rights are being terminated, relatives prevented from having kin placed with them, children cut off from siblings. Many of these calls involve appeals worth filing. But we turn most of them away because we simply don’t have the capacity to take them.

Every case we can’t take is a family that gets no one. Not a less experienced lawyer. Not a less thorough brief. No one. The appeal goes unfiled. The arguments go unmade. The family navigates the appellate process alone, or more often, doesn’t navigate it at all.

For most of my career, the math has been simple and brutal. Each case takes what it takes. There are only so many hours in a day. I can take on a few more cases, or I can do better work on the ones I have, but I can’t do both. And meanwhile, families keep losing.

This is the tension that has defined my professional life: the gap between the suffering I can see and the help I can offer.

A Different Question

And then the world changed.

I’ve started asking myself a question that I think every lawyer should be asking:

How might I use new AI tools to respond to human suffering in a way I never could have before?

Not “is this technology good enough?”

Not “does it meet our professional standards?”

Not “what are the risks?”

Those are fine questions, but they’re the wrong starting point. They begin with skepticism. They begin with protection of the status quo. And while we’re busy interrogating whether the technology is ready for us, families are losing their children with no one to fight for them.

The question I want to start with is different. It begins with the suffering, not the tool. It begins with the parent who just lost everything and the empty chair where their lawyer should be sitting. And it asks: what am I willing to do about it?

For me, the answer has started to take shape. I’m trying to build AI-assisted workflows for my appellate practice. It starts with tools that review transcripts and produce initial notes, so that when I sit down with a record, I’m editing and refining rather than starting from scratch. It might involve creating first-draft generators trained on my own writing – years of statements of fact, argument sections, appellate briefs – so the scaffolding is in place before I begin the real work of advocacy.

We should be teaching every lawyer how to do this.

Does this mean I’m no longer deeply engaged with my cases? No. I still read every page. I still make every strategic decision. I still write the final version of every brief that goes out with my name on it.

What it means is that the hours I used to spend on mechanical tasks – transcribing notes, formatting documents, building the scaffolding of a brief – are now hours I can spend on judgment, strategy, and advocacy. This is the work that actually requires a human being.

And it might mean I can say yes to cases I would have otherwise had to turn down.

The Mindset That Matters

I know what the objections sound like. I’ve heard them. AI makes mistakes. It hallucinates. It can’t replace the judgment of an experienced attorney. First drafts generated by a machine lack the depth of ones crafted by a lawyer who has lived with a case.

All of that might be true. And none of it matters to the parent who just lost their child and has no one willing to file an appeal.

The legal community has spent the last few years engaging in theoretical debates about whether AI is good enough. Whether it meets our high standards. Whether it threatens our professional identity.

Meanwhile, families are being permanently separated, and the only response our system can muster is “I don’t get paid” and “I don’t have time.”

It’s time to shift the default.

What if, instead of starting from skepticism, every lawyer who encounters these tools started by asking: how might this help me reach someone I can’t currently reach?

What if the first instinct wasn’t to catalog the risks, but to imagine the possibilities?

I’m not asking anyone to be reckless. I’m asking for a change in posture. Instead of demanding that AI prove itself before we’ll engage with it, what if we engaged with it precisely because the stakes of not engaging are so high? What if the burden of proof shifted – from “show me this technology is safe enough to use” to “show me you’ve seriously tried to use this technology to help people who are suffering”?

Think about what that mindset could unlock. If every lawyer asked that question earnestly, how many more families might get help? How many motions might get filed? How many appeals might get briefed? How many people who currently navigate the system alone might have someone in their corner?

We will never know if we refuse to try.

The Real Question

So here is the question I want to put to everyone in the legal profession who has a strong opinion about AI:

When someone is suffering and no one will help, what are you willing to do about it?

Are you willing to use a tool that’s imperfect if it means you can reach people who would otherwise get nothing? Are you willing to let a machine handle the scaffolding so you can focus on the substance? Are you willing to trade the comfort of your current workflow for the possibility of serving more people?

Or will you keep holding out for perfect while families lose everything?

Learning how to use the technology isn’t the hard part. The hard part is choosing to stay in the discomfort of knowing that people are hurting and demanding of yourself that you do something about it.

That’s not an AI question. It’s a human one.

But AI might be the tool that makes the human answer possible.