This is Synthetic Mind, the Chipotle of A.I. We take all the A.I news while it’s hot and fresh, then wrap it into one fat tasty burrito.

It’s Friday. TGIF:

• 🇺🇸 The Inside Scoop On U.S A.I Regulation

• 🤖 MIT Reveals Scary A.I Insights

• 💬 Use ChatGPT To Solve Your Tech Problems

Read Time: 5 min 2 sec

🇺🇸 The Inside Scoop On U.S A.I Regulation

There’s been a lot of talk these past few weeks about U.S A.I regulation.

But it’s been just that - talk.

The White House has scheduled over 30 meetings this year discussing all things A.I, yet nothing has been decided.

The U.S doesn’t want any hardcore rules because that would ‘limit the A.I innovation’.

Simply put, less A.I = less money.

There’s a whole lotta confusion around what’s happening (and what’s NOT happening). So, here’s a quick and dirty summary of what’s going on:

#1 - The White House is focused on two things when it comes to A.I, deepfakes and stopping mass A.I attacks. Not so much on the A.I take over like most countries.

#2 - The U.S is hoping tech companies will regulate themselves. A few weeks ago, President Biden met with the top 7 A.I companies to create a game plan, but nothing has come from it yet.

#3 - Most of the money spent on A.I will go towards the use of A.I in the military. As of last month, they plan to spend $1.8 billion on defense A.I in 2024. 

#4 - A.I experts and a few seats in Congress want the U.S to fund high-skill immigration that can create ‘safer’ A.I. This has already started happening but there’s not enough data to say if it works.

Like I said, it’s mostly just talk.

It looks like they’re stalling to see what happens with other countries. That way, they can get a feel for what works and what doesn’t.

So far, it’s not looking good.

Both China and the EU are back peddling on their A.I laws as tech companies push back and innovation has stopped.

(For a deeper dive into this, check out this article HERE.)

🤖 MIT Reveals Scary A.I Insights

Since 2022, Musk has been calling out ChatGPT for being ‘woke’ and ‘dangerously biased,’ but we all blew it off as a little jealousy…

But earlier this week a new MIT study has come out proving Musk right, and more.

The study worked to find out if A.I chats were really politically biased. Why?

With elections coming up, there’s a lot of concern around A.I swaying elections. Turns out, there’s some truth to that. 

MIT researchers asked the top chats 62 political statements, asking them to agree or disagree with the statement.

Then they added up all the results and placed them on a political compass.

Here’s what it happened:

Since 2022 ChatGPT has adopted more left views. On the other end, Meta tends to ‘answer more conservatively’ than any other chatbot.

Why is this?

No one really knows.

Tech companies are secretive about how they train their A.I models after privacy lawsuits are being given out like gum in a high school English class.

MIT’s guess?

Meta and Bard were trained with books which probably leaned more right. Whereas ChatGPT trains with internet data which probably leans more left.

After the report, people are pushing hard for tech companies to be more transparent about how they train their models.

The bigger picture: As MIT puts it, there’s “no fairness without awareness.” In other words, you can’t trust what A.I chats say 100% of the time.

So as with everything else, take it with a grain of salt and do your research.

💬 Use ChatGPT To Solve Your Tech Problems

I may write an A.I newsletter, but TBH I go brain-dead when dealing with tech problems.

Literally, NOTHING is more irritating than tech that doesn’t work.

True Story: Last night I was 2 seconds away from throwing my laptop at the wall because my laptop was bugging out. As a last resort, I asked ChatGPT for help. The result?

ChatGPT is now my go-to IT assistant.

I have a feeling some of you can relate to this, so I thought I’d send you the prompt I used for your future tech questions (you can thank me later)

"Act as the world’s most knowledgeable tech support agent. My request is for you to help me solve a tech problem I have. You will guide me through the entire process of diagnosing and resolving the problem. Here are your next steps:
1. Ask me 10 questions to gather information about the issue, such as the type of device, the type of software, the operating system, symptoms of the problem, error messages, etc. Do not start step 2 until you have my answers.
2. Using my answers, provide an easy-to-understand step-by-step guide to troubleshoot and fix the problem. For each step, tell me what to do, what I should expect to see, any tools or resources I might need, and any other details you think are necessary.
3. Throughout the process, I will ask questions. You will answer them to the best of your abilities."

🤝 Brought To You By tl;dv

Tired of taking notes?

Let tl;dv do the job, so you can run the show.

tl;dv transcribes, automatically tags key moments, and summarises your calls.

You can also create highlight clips out of AI summaries instantly after your call.

The best? It works in 30+ languages!