Website security is a public health issue


By webs ·

Listen to webs read this article (8 min)

Chapters

This reading is an episode of The Iterative Leader podcast.

Listen on Apple Podcasts (opens in a new tab)Listen on Spotify (opens in a new tab)

Anytime I think a website was vibe coded, or someone asks what I think of a site's security, I run it through securityheaders.com (opens in a new tab) or Mozilla Observatory (opens in a new tab). I did that with someone's site the other day, and it got an F.

If you built your site with ai, yours could get the same grade without you knowing it. A gap in your site may not seem like it impacts you, but it does, because it puts your customers and the rest of the internet at risk. What happens on your site reaches everyone who visits it, the way one person's health reaches the people around them. That's why I think website security is a public health issue, and that we should treat it like one. It is one because attacks reach people's health. When ransomware hits a hospital, deaths among patients already admitted go up by about 20 to 35 percent (opens in a new tab), according to researchers who studied Medicare patients.

I'm not the first to make the case. Security researchers have compared the two since at least 2012 (opens in a new tab), some call it a public health crisis (opens in a new tab), and a 2026 paper (opens in a new tab) proposes a national cyber public health system. That work maps threats and fixes from one field onto the other. What I want to add is the part about changing hearts, and how we talk about security so people care about it. Below, I'll walk through what that F means, why caring about security matters as much as knowing about it, and how to check and fix your own site.

Your website is part of a web

Your website is just that, a site that's part of a web. It's wild, I know. We say "website" all the time and still forget that our little corner of the internet is connected to all the other little (or big) corners of the internet. When your site is compromised, the damage doesn't stay on your site.

Take that F. Among other things, it means anyone can load the site inside a frame on their own page and steal logins through clickjacking. Clickjacking is when visitors think they're clicking your page, and they're really clicking something hidden on top of it. The logins that get stolen belong to your customers.

Some attacks also spread from one account or site to the next, the way a cold moves through an office. In 2005, someone found a cross-site scripting flaw on MySpace, a gap that let him add his own code to the site. He planted a worm, code that copies itself, on his profile. Everyone who viewed his profile got the worm copied onto theirs, and everyone who viewed those profiles got it too. In less than a day, more than a million profiles had it, and MySpace went offline to clean it up (opens in a new tab).

Security doesn't always mean an attack from outside, either. It also covers keeping systems up and working, which security folks call availability. The recent FAA ground stop (opens in a new tab) traced instability in the air traffic communication system to a router in Arkansas. Two days later, more than 1,600 flights were still canceled. Also last week, an Outlook outage (opens in a new tab) that Microsoft traced to a network change ran across two workdays, drew more than 100,000 reports within an hour, and led New York City's Emergency Management Department to warn residents they might have trouble reaching city agencies. Neither was an attack from outside, and both show that security threats have real world impacts.

I think it's amazing what ai has opened up. So many people can now make their ideas come to life in a way that was previously out of reach. Most of them don't have the background to know the security, privacy, and accessibility issues that come with it, or how those issues put their customers, themselves, and the rest of the internet at risk.

Knowing about it isn't enough

Security takes educating minds, and it also takes changing hearts. You can know every best practice for guarding against a breach, and at some point the easy way to do something will also be the less secure way. If nobody has built a culture where security matters, all the education in the world goes out the window the moment the secure thing is inconvenient. People have to care enough to keep their corner of the internet clean and healthy.

I've seen this happen. On my first day on one of my first projects at Thoughtworks, I asked for the Wi-Fi password, and someone sent it to me in plain text over chat. It wasn't the most egregious error, and also small things like that are how hackers breach companies. I was appalled, and I also knew shaming wasn't going to help anyone. They sent it that way because nothing more convenient existed.

Telling them the rules wouldn't have been enough, because the convenient way would still win. They had to care, so I went to work on that.

I started with a joke. Whenever something like that happened, I'd say it breaks my security heart. The joke called out the insecure choice without shaming anyone, and it admitted the insecure way was the convenient one. Then I made the secure way the convenient way. I showed the team One-Time Secret (opens in a new tab), a free tool that turns a password into a link that works once and then disappears, so the password never sits in a chat history. A password manager account that allows sharing between users would work too. Over a few months, the team started choosing the secure way even when it took an extra step. They did it because they valued me as a teammate, because I kept showing them that the secure way was about their own integrity, and because we built systems to make the secure way more convenient (although that's not always an option). By the time new people rolled onto the project, a one-time link was how everyone got the Wi-Fi password.

Public health has worked on this same problem for a long time, and it calls the work risk communication (opens in a new tab). These are the tenets I borrow from it when I help you with your site:

  • Say what's known and what isn't, so trust builds over time.
  • Express empathy and validate your experience.
  • Give clear, manageable, and meaningful steps.
  • Use plain language.
  • Make the risk personally relevant.
  • Treat it as a dialog, listening to your questions instead of handing down instructions.

Empathy comes first. Nobody should be made fun of or made to feel crappy for not knowing, and there's nothing wrong with asking for help. With a learning sciences background, I meet you where you are and step you through the process, so you don't have to feel silly or self-conscious.

Your first step

Run your site through Mozilla Observatory (opens in a new tab). It's free, anyone can use it, and it grades your site and lists what to fix. Go through the recommendations one at a time. You can ask your ai to review the findings and update your site to guard against them.

For example, it might say you haven't set a content security policy. A content security policy sets the rules for what code is allowed to run on your site. It helps prevent cross-site scripting, the kind of attack behind the MySpace worm, by controlling which resources your site loads. It also restricts where your site can be embedded, which stops the clickjacking behind that F, and it can upgrade insecure requests to HTTPS.

I'm your tech Mary Poppins

If that last paragraph felt like a lot, you're not alone, and that's where I come in. I like the Mary Poppins approach. I want to partner with you, build up your skills, and then move on to someone else who needs me.

If you want to know what your findings mean, bring a question to my free office hours (opens in a new tab), Wednesdays at 8:30 am Pacific. To go deeper on your own site, book a one-on-one hour (opens in a new tab) with me for $225.

Thanks for reading.

If this raised a question about your own team or product, ask me.

More writing lives on The Iterative Leader.