When Does Helping Violate Do No Harm?
Another post where I violate the ironclad growth advice of the all-knowing growth people — have a strong take! be a clear brand! — and ruminate a bit on something I do not have a clear opinion about: whether or not some help is better than no help.
On the face of it, that sounds stupid. Of course, some help is better than no help. Don’t let the perfect be the enemy of the good and other cliches! If you can do something to help, then it is better than doing nothing to help. But in the specific case at hand, I am not really sure that it is. Specifically, I am discussing AI health, education, and therapy assistants. A recent Wired article point out that we are rapidly approaching a system where rich people get humans to help with their health care and education while the less fortunate get only AI and other automation. As one creator of these tools put it, they are not as good as human help, but they are better than nothing.
I am not sure that they are. Take the example of how AI is used in schools. In one wealthy school it is an augment to at least 45 minutes of individual attention a week, plus regular group work, and daily attention from a human teacher who is an expert in the subject. In a poor school in Mississippi, students who need help beyond what the machine can provide have to wait for a teacher from another city to find time for them. Now, one could argue that poor students still benefit — they are getting some education, are they not? But are they?
Everyone who is learning something is going to get stuck. Having no access to teachers who can assist means that any education is, at best, incomplete. At worst it is harmful, leaving students with a flawed understanding of the material or a sense that they know more than they do. It could also turn off students to the subject completely — why bother when no one is going to help you?
Similar issues arise with therapy. Human connection is superior to inhuman connection. We are social creatures — studies show that something as simple as talking to your barista improves your mental health. But, as the article notes, the drive for efficiency in health care means that doctors spend less and less time getting patients to open up to them since they are driven to see and more and more in order to survive financially. AI and automation might alleviate some of that time pressure for the non-rich, but machines are not now and are not likely ever to be real replacements for actual human connection.
Which brings us back to whether or not these are better than nothing. On the one hand, there is no doubt that they fill a gap. Even if they fill the gap imperfectly, some help is always better than none, right? But money and time spent on building out these systems could be money and time spent on advocating for getting every human being the human driven services they deserve as human beings. If we allow these systems to become entrenched, if we allow good enough to replace good, then we do real harm to the people that we are supposed to be helping.
On the other hand, getting people real services is an uphill battle that will not be won anytime soon. Not accepting help, even flawed help, means that some people get no help at all in the interregnum. It is easy to be worried about the long term when you aren’t the one suffering in the short term.
I wish I had an easy answer for you, but, obviously, I do not. No one died and left me God (the Blackhawks wouldn’t be, well, this, if someone had, for starters). All I have is maybes and analogies. And the more I think about it, the more I am afraid that we will be looking at a poorer version of the ACA.
When the ACA was trying to get passed, the left gave up on Medicare for all in order to get something passed and was willing to settle on two compromises — lower the Medicare age to 55 and a publicly run medical plan that people on the exchanges could buy into. Both were destroyed by Senators who were in the pocket of health insurance companies and/or had made their public persona little more than hippie punching. In two subsequent cycles of Democratic control, nothing was done to add a public option or decrease the Medicare enrollment age. The status quo must be defended, it seems, not improved upon.
The ACA is not good health care legislation. It still leaves far too many people paying far too much out of pocket and gives insurance companies too much control over people’s lives. The rage at the UnitedHealthCare CEO who was recently assassinated is testament to the depth of that anger. However, it is much better than the world we used to live in. Protection for people with existing conditions, requirements that insurance spend most of its money on care not profits, requirements that some things be free and available to all, subsidies for some people who need help — all of these things are worth defending. We can live with this system until we can force a better one.
The likelihood is that we will come to accept these inhuman systems of care and education as part of life, the new status quo. The question then becomes are they like the ACA — flawed but can be lived with? Or are they so bad that they actively harm people and their existence does more damage than they alleviate? I don’t have a clean answer. I suspect, given the technocratic and Silicon Valley worship of efficiency above all else, that these systems will be harmful above all else. But this is not an easy or clear question. And the fact that sometimes deeply flawed systems, like the ACA, are still vast improvements, complicates the questions.
Despite the eternal cry of the centrists, the good sometimes really is the enemy of the perfect. The trick, unfortunately, is that sometimes it is not. It is up to us to not rely on hoary cliches and comforting bullshit passing itself off as wisdom and interrogate the systems that come in contact with our lives deeply and honestly. Anything less and we risk leaving people in harm’s way.
And that is neither good nor perfect.