Trust, Safety, & Security.

Archived from occybyte.com/resources · 2024-09-13

Something that has been on my mind lately is Tech Policy and AI Trust, Safety and Security. Something sparked in my mind in 2023 when I was learning about AI, and then I decided to start learning as much as I can about it.

AI is now so intrinisicly invovled in our daily lives - and it has been for years, but now with the advent of generative AI - which data: ours, yours, theirs, et cetera, it is all used to train these models. They are linear regression machines that have access to all the knowledge in the world. Data is used to fuel technologies that have caused a paradigm shift in how we percieve computing and art.

I’ve always been a tech enthusiast, and I’ve always been interested in Defense in some aspect. I’m still budding in my career in general through game design, through writing and through AI and cybersecurity. I’ve always been an analytical person and when I was younger I loved creating fake documents.

I had a curiosity about computers and because I didn’t have a computer that could play games when I was younger, I would play with notepad and make batch files, learn about the registry and I’ve always been fascinated with the command line interfaces. I was one of those kids that would tinker with the settings in Windows XP and I was always curious about how things worked.

So, naturally, one of the first games I made on my own was a point and click called VoidInstaller. It was based on what I’ve been writing in the Terminus, Godliving and Godless series basically. Void.exe was simply a game where the player had to click ‘boxes’ that were viruses and they earned a currency, which they could use to buy shields and they also had to manage their hunger. Having a lot of viruses on the screen would cause the system integrity to go down until it crashed.

Simple concept, simple game. Not fun, but it was an idea that I got done with a distinct theme and an idea to execute.

Cybersecurity has been something that I was interested in since I was in middle school. It’s funny, I applied to the University of Louisville back in 2013 or so, and I actually won a scholarship to the Speed School of Engineering. 10k buuuut, I was too depressed - suicidally depressed. Thus I didn’t go. I do have a lot of things I regret, that is one. Not joining the U.S. Air Force was another.

Policies, analyzing and going over data has been things that have invigorated me.

RISKS IN TECH POLICY

One of the biggest things that I have identified is that a lot of knowledge, talent and processes in the tech indusrtial are centralized to only a few locations inside of the United States. This is not good for a multitude of reasons, but one of the main ones is that it is not diverse.

The State of California is an interesting example - it’s a doubled-edged sword. Silicon Valley are innovators and pushers, the San Francisco Bay Area is where a lot of personnel are located. All the funding and talent is there due to it being a hub. But it’s causing brain drain to other areas that would benefit from that knowledge and those people.

As a result, the San Francisco Bay Area is highly susceptible to group-think, there is an high risk factor in making people need to move to the Bay Area and it’s a high cost of living area. As a result, in order to maintain their coveted positions (thus maintaining their income, benefits and lifestyles) there is an implicit bias to conform to the current trends. This isn’t good for culture, nor is it good for innovation.

And even more concerning this is that because companies here are coveted to join, this means that there is sharing of information of hiring procedures and tests that would otherwise deny someone. For example, a hiring assessment based on how the person works. There is likely a score or an algorithm that is used to determine if someone is a good fit for the company or not. Which is fine, but it can be gamed and people will game the system to get into these companies; even more importantly, they will share this information to help others game the system.

As a result, people are are overfitting themselves to gain an edge of competition. This is not good for innovation. Another example is with leetcode, which while good, it only forces people to memorize via ROT Memorization. This is not good for innovation, not good for critical thinking and especially creative problem solving. Which I’m mentioning because these are also generally timed.

Competitive programming is like a resturant hiring someone based on their speed-stacking ability. It’s not a good way to determine if someone is good for the team - while there are other tests - especially behaviorial and cultueral.Leetcode is a way to test if someone can make an algorithm give an output.

For example, Google? Being innovative? What they should do is group programming as a test with a ‘cohort’ of people.

While we do have Standards and Code of Ethics, this is more for policy and law.