When weak decisions repeat, people stop seeing them as decisions at all.
Harm does not arrive with a dramatic speech or a flashing warning sign. Sometimes it arrives disguised as routine.
A new rule, a missing explanation, or a form nobody can challenge. It could even be a policy that people are told to accept because it is already moving, has already been approved, or is disguised as something else entirely. Some of those things could have been decisions that would have shocked people a year or two ago. However, once they get described as necessary or efficient (or “just how things work”), normalization steps in.
After all, once harm becomes ordinary, people stop arguing or scrutinizing whether it is happening, if it has already happened, or if it’s worth mentioning at all.
This can look different across the board. It could pertain to data privacy or data protection; the reality or discourse around algorithmic harm versus algorithmic discrimination; due process and fundamental rights; and the list goes on. P.S. I wrote more about the difference between cybersecurity and data privacy in an earlier post.
This post is about those kinds of shifts and how they impact entire systems.
So, what gets something normalized?
As with most things in our day-to-day reality, there’s a timing component. More specifically, think about how long it takes to market a new product, how long rebranding initiatives last, or shifts in company culture after you get a new CEO. Acceptance is not overnight. It takes time. It is just about impossible for anything new to become normal all at once, because change happens in pieces. The only exception could be someone trying a Dairy Queen cake for the first time and implementing it into their weekly LIFE (I haven’t had lunch yet).
Anyway, when a harmful structure becomes so common that people stop talking about it in terms of (not) having a choice or influence on something, because they fear for their job. That’s how an abuse of power gets dressed up as policy. It’s also how a loss of access gets framed as a trade-off or treated as administrative complexity. That is also part of who gets a say before decisions harden into structure. Then, over time, the (extra)ordinary starts sounding procedural.
Why normalization becomes structural.
Normalization isn’t just cultural. It is structural because it depends on repetition and silence that create overwhelm and confusion, making people feel as if pushing back would cost too much.
Ultimately, once people stop asking whether something should be happening, they start focusing only on how to survive it. By then, who absorbs the risk is rarely the person with the most power.
This is one way creeping authoritarianism works.
If history has taught us anything, it is that creeping authoritarianism normalizes through quiet pressure, and it is never something that a society intentionally wants. Things like plain language being scrutinized and changed into something seemingly “cleaner,” not clearer. At the same time, public conduct gets rougher. Accountability gets harder to enforce. People are pushed to accept more and more unethical directives, less transparency, weaker oversight, and more fear around dissent than they would have tolerated before.
Sound familiar? You can see parts of that in the U.S. right now.
Creeping authoritarianism examples across sectors
For legal folks
Legal scholars and former judges told the Associated Press that the current administration has shown an extraordinary pattern of violating court rulings (across a wide range of cases). There is also mention that the Justice Department’s posture has been unusually combative when challenged. That matters because when compliance with court orders starts looking optional, we are all being asked to treat checks and balances as negotiable. I’m no lawyer, but this seems to be both a priori and a posteriori knowledge, by the way.
What it looks like in higher education
You can see it in higher education, too. Human Rights Watch described academic freedom in the U.S. as under threat, citing funding pressures and ideological attacks on universities, usually tied to campus protests and speech.
In healthcare
In healthcare, the administration just targeted California by cutting $1 billion in Medicaid funding because they don’t support this regime’s stance on “intolerance.” There are many more examples from this sector, too.
Tech sector
In tech, think about the discourse versus reality of things like algorithmic harm versus algorithmic discrimination. Arvind Narayanan argues that the fairness frame can miss broader harms from algorithmic decision-making when people focus too narrowly on discrimination alone. Discrimination is a piece of overall algorithmic harm, but it is far from the entirety of it. That does not mean to justify discrimination by any stretch of the imagination. I am pointing it out because everyone seems to love algorithms nowadays, and that distinction is important. There are no panaceas.
My point is, even if someone wants to debate, the pattern is familiar: use institutional pressure to narrow what can be said, taught, challenged, or publicly defended.
That is why it is called “creeping authoritarianism.” It does not need to look identical in every sector to still be part of the same pattern. At the same time, all of our systems are connected, whether directly or indirectly.
The problem is not just the policy…it is the adjustment.
That is where normalization gets dangerous.
The damage is not only that something bad happened, but that people are pushed to build their routines around it. Then, after some time, someone new comes in and views the setup as “normal.” Why? Because to them, it has been that way since they got there. How can they question otherwise? At that point, it is already embedded into the workflows or protocols.
That is how a system teaches acceptance.
Perhaps, it is also why people say “history repeats itself” or “hindsight is 20-20.”
Opaque systems help bad norms settle in.
This is one reason I keep coming back to infrastructure, documentation, and accountability rather than treating them as separate risk management factors or back-office details.
Data & Society’s primer on algorithmic accountability notes that there are still too few protections and auditing standards for these types of systems. Although this was published in 2018 (I just learned that there’s a lack of research in this area), you can read that primer here: Algorithmic Accountability: A Primer (as of 6/25/26 – the link is broken). Specifically, systems that affect housing, healthcare, hiring, social services, and other aspects of daily life. It also argues that accountability should include a mechanism for redress when harm occurs. That matters here because when the “logic” is “hidden,” the public is left adjusting to outcomes it cannot interpret. From another perspective, this is also part of what gets tracked and what quietly disappears.
There’s also a Stanford case study on algorithmic decision-making and accountability that makes a related point from the public sector. It opens with examples from NYC, where agencies relied on formulas or algorithms to allocate services, school placements, and other outcomes, while the people affected could not really see or challenge how those systems worked. That is exactly the kind of environment where harmful decisions can become routine faster than people realize.
If you would like more examples, here are some case studies from Stanford: https://ethicsinsociety.stanford.edu/tech-ethics/education-programs/case-studies
For the HR folks
A Princeton hiring case shows how a system built for efficiency can slowly displace other values like diversity, equity, and judgment. You can read the case study here: Hiring by Machine. That example is not only about hiring software; it shows us that once a system seems fast, scalable, and good enough, institutions start trusting it. Ultimately, in cases like this, the system’s value judgments stop looking like value judgments at all. They just look operational.
That is normalization in the business world. And that type of normalization is not neutral, because it is not designed that way.
To be fair, not every routine is neutral.
This is where I think people get tripped up. They assume that routine means ordinary, and that ordinary means acceptable.
No.
Some routines are just habits. Others are signals that a harmful structure has settled in, and people have started organizing themselves around it. For example, if workers feel they must remain silent to protect access to their paycheck, that is not neutral. The same goes for communities expected to accept more surveillance, more secrecy, or more intimidation as the price of safety, which is not neutral. P.S. Facial recognition will not reduce the number of times a wrong person is arrested. It is also not neutral when institutions start treating constitutional limits, due process, or public accountability as obstacles rather than safeguards. Additionally, that is where selective enforcement starts to show itself.
What do we do with that?
So, collectively, we must avoid saying and believing “that is just how things work now.” We need to believe our own eyes and intuition.
Second, I think we have to either begin now or continue naming normalization more clearly, because not everything harmful looks like a crisis in real time.
Third, I think we have to get better at noticing the ordinary parts of our everyday lives, become familiar with systems thinking at a local level, and understand the differences between the two.
And finally, I do not think this is only a matter for elected officials, lawyers, or public figures. It is part of how to defend rights from where you are. For example, this shows up in how people document things (shout-out to the constitutional observers, movement lawyers and legal workers out there), whether we continue to question a bad process, and if we can connect the dots.
That is slower work, but it is real work, and it is the kind of work that keeps us free.
At the end of the day
Don’t be afraid to set boundaries, because how you present and show up are shining examples for others. Many systems do not become dangerous only when they break, but become dangerous when people get used to carrying what should have been questioned much earlier.
That is also why I keep asking difficult questions in my work: once harmful patterns become routine, it becomes much easier for institutions to keep expanding what people are expected to accept. That is not simply a policy, something too technical, or solely an operational problem; it is a societal problem.
Individually, we are powerful, but collectively, what we normalize becomes our reality.
That is what we should consider normal. Not the opposite.