😺 Politicians return to college

PLUS: Chinese company sued for being "ageist"

Welcome, humans.

You know when your computer overheats and that noisy fan kicks in? Yeah, get used to that … cause we’re coming with the heat this week. 🔥

Here’s what’s going on in AI today:

US policymakers attended an AI crash course at Stanford.
Thousands of hackers tried to find AI weaknesses this weekend.
A company was sued for using AI to discriminate in hiring.
AIs can pass reCAPTCHA tests…

Sponsor The Neuron | New Here? Subscribe!

The US Is Finally Acting On AI.

Washington Post

Last week, dozens of US regulators and politicians gathered for an AI crash course at Stanford. Why?

To figure out how-the-hell to handle AI.

So far, the US has been stuck in a “learning” versus “doing” stage, but it’s beginning to get its sh*t together.

The first objective is to defend the US:

Last week, Biden proposed a ban on US investments in Chinese tech companies, which includes AI.
His team also launched an “AI Cyber Challenge” with $20 million in prizes for AI solutions that defend US national security.

But that’s the easy part.

The hard part is managing AI … inside the US. We’re referring to:

Deepfakes.
Copyright standards.
Superintelligent AI … aka AGI … aka potentially the end of the world?!

Why it matters: The White House took baby steps this weekend by contributing to the DEF CON conference, where thousands of hackers attempted to find weaknesses in AI models from OpenAI, Anthropic, & more.

This process is also known as “red-teaming.” For the curious, a recent study tested 6,387 prompts and identified two that consistently trick an AI into responding to questions it shouldn't.

FROM OUR PARTNERS

You’ll regret not attending this AI conference

SF is back, and it’s the place to be if you want to build AI (said by a New Yorker).

Here’s an opportunity to see be in the room where the top innovation in AI happens:

The AI Conference — a melting pot of researchers and engineers working hand-in-hand to understand challenges and build solutions.

Collaborate on projects with brains from Anthropic, Nvidia & Hugging Face. Instant career boost?

As a loyal Neuron reader, you’re in luck: The AI Conference is offering our readers a 15% discount (yup, we deliver). Just use discount code “Neuron 15”.

But you gotta move fast: tickets are running out! Secure your spot before the opportunity disappears!

Companies That Use AI In Hiring Need To Tread Lightly.

Made using Midjourney!

Last month, NYC passed a law requiring companies to pass a bias audit before using AI algorithms in hiring.

This makes sense. Algorithms are created by humans. Humans are subject to biases. Thus, algorithms will also be biased.

(This might explain why only 7% of Americans favor AI making the final call in hiring.)

Just last week, a Chinese tutoring business paid $365,000 in settlements after it was caught using AI to unfairly filter out older job applicants. Don’t disrespect your elders!

Why it matters: Using algorithms to vet applicants isn't new; recruiters have been doing this for years. AI will make this process even more effective.

But companies looking to hire via AI need to be extra careful that they’re doing so in a fair and transparent manner.

Around the Horn 🦄

AI detectors are very easy to fool with additional prompting and are biased against non-native English speakers.
Here’s an ultimate list of generative AI resources (reading, projects, tools).
Robots are actually better than humans at “I am not a robot” reCAPTCHA tests.
A majority of companies have considered or already instituted a ban on generative AI at work, according to a study of 2,000 IT execs.