And organisations fail at security when security and compliance is only considered important by that one teamnn

Security requires the whole business to buy in. And it requires processes that allow people to get shit done without people resorting to shadow IT; thus working around that one team.

So the GPs point still works.

In a high-functioning security and compliance team they tend to say “no” only in really dire circumstances. Good security and compliance teams spend a lot of time asking exactly what it is the people are really trying to do and then find a way to say “yes, and…” as in “yes you can and here’s how you do it without compromising security/breaking the law etc”. And as a sibling said, orgs fail at security when they make security only the infosec teams’ job.

This is also why I said “whose only job”. In a good org, the security team doesn’t only say no to devsecops requests, they also do trainings to skill up other teams, keep the network secure, proactively seek out and understand external threats, work with external vendors etc etc …

I think the key part is about aligning incentives and outcomes - if your security and compliance departments are only judged on "were we breached, did we pass our audits" then there's a risk they won't weigh the tradeoffs associated with the decisions around controls implemented to achieve those outcomes.

To use a ridiculous extreme you can't breach a web app that isn't exposed to the internet, but the users can't access it either.

If you can connect/balance those goals to other metrics around cost and productivity, usability, and a realistic threat model, as guardrails then you incentatize cross-team collaboration to achieve the shared org outcomes.

The structure is untenable when the work of one team is to say "no" to other teams, when these teams are not asking.

A good infrastructure team would seek a competent security review that would say "no" to problematic things before an intruder says "aha" to them. If feedback from the ethics team is not sought, nobody is going to heed its opinion anyway.

SRE is a 'say no to powerful people' job as well. I think for this to work the leadership needs to show support for it. The friction is still there, but in more tolerable areas. I bet for ethics this isn't the case at OAI, but everyone values security and stability.

For an SRE there can be more directed hate received from the junior employees, that want to release new features they developed. Especially because there is less accountability across orgs. Security is an interesting one because it seems to have less of this friction, maybe because it's more clear cut what is an issue.

During my years at big corp we did tons of work just to go around the law department. While for core business it made absolute sense to have lawyers involved in daily business we were satellite office decoupled from core business. Local management failed to communicate that and we were stalked by lawyers from main office. Pretty sure others do the same with compliance, law, ethics departments. Just work around to get things done instead of stuck for weeks in stupid meetings.

Security/compliance have the external hammer: if they fail, your org will have to pay fines and someone may end up being criminally liable (depends on country).

The ethics department; if they fail, there may be some negative journalism, but who which AI company has positive journalism these days? There's no external hammer for ethics.

Security and compliance tends to be a lot less subjective than ethics. Not saying it’s objective, but there’s a huge difference.

The downside is it can often feel like a box-checking exercise than actual security or compliance, but “you need 2FA” is less debatable than, say, AI and copyright.

Markets solve this by pricing risk. Ideally you would have some kind of notion of selling insurance internally and track things. But ultimately existential risk is hard to negotiate from the inside. And companies are supposed to go bust or succeed. They are not really the same as a population trying to survive forever. At least that is one take.

There are at least some penalties for failed security and compliance. Ethics is the first victim of greed.

> exactly how security and compliance

Not everywhere. I go out of my way to assist teams to achieve a secure outcome with less effort.

Things like: “instead of admin access to the production servers the devs can have fully automated deployment pipelines combined with OpenTelemetry for observability so they don’t have to spend half the day scrolling through gigabytes of logs.”

That’s more secure and and more better.

Nobody had to be told “no”.

Similarly, I replace key store access with secret-less managed identity, etc.