Biden’s Executive Order Puts Civil Rights In The Middle Of The AI Regulation Discussion

TPM Article Topics
1

This article is part of TPM Cafe, TPM’s home for opinion and news analysis. It was originally published at The Conversation.

This is a companion discussion topic for the original entry at https://talkingpointsmemo.com/?p=1473749
3 Likes
2

Nothing in there about how to stop Skynet from killing us all or turning us into batteries.

3 Likes
3

Garbage in and garbage out. I have looked at this issue and I don’t see anything about AI that makes its solutions inherently better than human solutions. Faster yes, but better no. At the bottom of all AI is a human being or a group of human beings putting algorithms together. In doing that they bring their human prejudices to the table. The AI they produce is going to reflect their human prejudices.

I have been thinking about the software company I am associated with. When we built our company we intentionally built it to be as inclusive as it could be. That was the intentional work of our executives and our HR professionals. Their is nothing to stop another company from incorporating their own hatred and prejudices in the work they do.

If you watch the AI sausage being prepard you rapidly come to the conclusion that Skynet is a very long way off. The best AI apes self-awareness but it really isn’t. It is just a different way to make sense of large amounts of data. It does reinject mystery into software development but it isn’t majic.

12 Likes
4

You are spot on target. AI depends on its data source. If that source is flawed in any way, your AI is flawed.

6 Likes
5

With generative AI it’s a combination of flawed data and flawed humans training the model’s output. Racial and other bias can creep in from both directions. It’s going to be very difficult to weed out completely because the AI will just reflect the bias of the society that builds it.

I read the fact sheet on the “Safe, Secure, and Trustworthy AI” executive order, but it seems mostly just aspirational fluff, without any teeth to stop the immense commercial push to get generative AI on everyone’s phones and into everyone’s lives.

Bill Gates wrote a post on his blog praising the coming of Personal AI Assistants, and I can just imagine how companies are drooling at the prospect of being able to scrape that personal info as a marketing tool. The privacy aspect of AI collecting personal data as your best buddy living on your phone isn’t being looked at hard enough. It’s a potential gold mine that goes far beyond what Facebook and other social media can do now with your data.

11 Likes
6

I am less worried about Skynet and more worried about Soylent Green.

2 Likes
7

I am a hardware guy, but know enough about S/W to understand the difference. Snark aside, the real concerns are human prejudice, as you already mentioned, and general sloppiness of the sort that causes automated systems to do dumb stuff like fly airplanes into the ground.

12 Likes
8

Old, old computer joke, playing off the hardware-vs-software divide

What are the three most dangerous things in computing?

  1. A programmer with a soldering iron

  2. A hardware tech with a patch tape (raise your hand if you know what that is)

  3. A user with an idea

6 Likes
9

I have learned always (mostly) to ask the wife before doing or saying something. AI presumes to compete with her authority. It’s not going to be pretty for AI.

9 Likes
10

I agree about spousal authority. Mrs darr and i have been together a very long time. That took practice from both of us. AI would be in serious trouble if it thought to piss her off.

1 Like
11

And faster means cheaper. AI tutors will be worse than human ones, along with being, well, inhuman, but they’ll be a lot cheaper. AI will put money in the pockets of big capitalists, and drain money from the public sector while corroding it. All the rest is advertising, seems to me.

2 Likes
12

That’s all this is about. The tech billionaires all racing to become the world’s first trillionaire. That’s all they really give a fuck about. They’d outsource, automate and eliminate every single human being at every company if they could without the slightest idea of the dystopian hellscape they’re creating by hollowing out the entire economy with generations of unnecessary human beings.

3 Likes
13

Some smart AI expert put it like this: The problem isn’t that AI will become super-intelligent and take over the world. The problem is that it will remain stupid, and it already has taken over the world.

4 Likes