What's up? You're reading Inclined AI. We joined Threads the other day but can’t think of what to call posts since “threads” feels a bit redundant, right?

Anyways, here's our thread of news:

• G/O Media is not having a great time

• Superalignment: a new initiative from OpenAI

• AI deserves a soul

• AI hiring software takes a hit in New York

Be a sponsor of our newsletter.

DON’T READ ANY BOT ARTICLES; THEY’RE TRASH

Hey there! It’s me, a biased human. I’m writing to tell you all about the lousy job AI journalists are doing at their job. Yes, I write daily and am affected by this news, but that doesn’t detract from the facts.

Last week, G/O Media told staff that an AI bot would start writing articles for their publications. At face value, that’s not a big deal. Other outlets are following that trend.

But all this AI content tends to be wrong.

The latest snafu involves io9 and an article that G/O higher-ups published without a deputy editor’s permission. The report was the first written by their AI model and centered around a list of Star Wars movies and tv shows.

The only problem is the bot littered the article with factual errors and grammar issues. Other than that, the piece was excellent!

It’s the latest failure in a string of problems. CNET, another publisher that uses AI, had to go back and edit a ton of AI pieces after their bot failed to stick to the facts.

All this goes to show that writing is more complicated than we realize.

It’s clear that G/O Media wanted to use AI to help their SEO rankings and get more eyeballs on their sites. I get it. No one wants to fall behind in a game that involves earth-shattering technology.

However, execution means everything here. Treat the bot like an intern. Would you let a fresh-faced writer publish a piece without a once-over? Hell no. Why would you treat a chatbot any differently?

A simple piece about Star Wars content doesn’t seem like a big deal, but the implications are massive.

Imagine a world where AI breaks every news story and decides what facts to include and omit. See, whether we care to admit it or not, journalists, in all forms, are the bastions of truth in our society.

You wouldn’t trust a machine to diagnose you without human assistance, so why would you trust a machine to tell you the truth?

We still need a human in the process; this most recent spat proves that.

Check Out Our Sunday Edition

You can dive into more AI news and topics with us every week by subscribing to our premium edition.

We’ve written about the following:

• Here Is The Work AI Will Replace Soon

• 8 Role Prompts that Still Work on ChatGPT

• 10 AI Tools You’ll Actually Use

If you’re not already subscribed, that’s okay. We’re offering a free 7-day trial so that you can read this one. That’s how excited I am to post it.

Claim the offer below to support our writing 🙂

WE HAVE 4 YEARS TO SOLVE THE SUPERINTELLIGENCE ALIGNMENT PROBLEM

OpenAI is assembling a team of top researchers and engineers to prepare for the next phase of AI development. Actually, it’s more like the next-next development phase, but who’s counting.

By now, you know about Artificial General Intelligence (AGI), the idea that AI will scale to match our intelligence level across mediums. ChatGPT is a step towards that, but its struggles with math are one example of how it doesn’t quite meet the requirements.

We’re not talking about AGI, though.

OpenAI—specifically, Ilya Sutskever (cofounder and Chief Scientist of OpenAI) and Jan Leike (Head of Alignment)—worry about Superintelligence, a form of AI that surpasses human understanding in every way.

Here’s the rub: we are more intelligent than current models. We can teach them and align them with our ideals (most of the time). OpenAI made ChatGPT using Reinforcement Learning through Human Feedback (RLHF), but humans won’t cut it forever.

The scale that Superintelligence lives at is unimaginable.

We will struggle to align an AI that is smarter than us. The AI can lie, ignore parameters, and trick the meat sack on the other side of the conversation. OpenAI wants to jump ahead of this problem.

That’s why they’re recruiting people for a project that focuses on Superalignment. All they’re doing is finding new ways to train AI that scale with the models; it’s that simple.

But the solutions aren’t. Executing this concept will take brainpower and tons of it. As the headline mentioned, OpenAI expects Superintelligence to arrive in a decade or so. They want to find a way to align it in the next four years.

Will they do it? Time will tell. However, seeing the company focus on its original mission and return to the big goals again is refreshing. ChatGPT features are fun, but we’re here for the singularity moment.

Quick Nuggets

🫵 Give AI a soul to solve the issues around AGI

🗣️ AI can help with communication in language classrooms

😬 The man who tried to kill the queen claims an AI chatbot told him to do it

🔞 Lovense adds a new AI chatbot to talk to you while you use their product

💼 There’s a clear match between AI and lawyers when it comes to communication

📄 AI hiring software needs to prove its unbias opinion in the state of New York

♻️ AI affects climate change in ways we don’t even consider

🌩️ The media storm that swept up Twitter and Reddit is caused by AI tearing up the old web

🔥 Fresh Products

• VEED - generate pro-level AI videos using text (link)

• FlutterFlow - build an app w/ AI by your side (link)

• Whimsical - translate ideas into flowcharts in seconds (link)

• Convertfiles - convert your images online for free (link)

• plugin surf - ChatGPT plugins, sorted (link)

• Sauce - automate customer insights w/ AI product feedback (link)

• GPT-Migrate - migrate a codebase from one language to another language (link)

• Lingoedit - simplify translation & editing (link)

Good Content, Sesame Riots

The War on Cookies was a bad policy decision, and history reflects that.

❓ Test Your Knowledge

Put your AI knowledge to the test with our quick quiz!

Reply with your answer, and check out the next issue for the correct response!

Yesterday’s Answer: B) Transformer

_{- That’s it for today. I hope you enjoyed the latest edition of inclined.ai - Davis.}