What's up? You're reading Inclined AI. You made it to Friday! Give yourself a round of applause.

Here's how to celebrate:

• A breakdown of OpenAI’s approach to safety

• Examine Google’s CEO and his latest promises

• Microsoft adds even more AI integrations

• AI video is reaching a tipping point

Be a sponsor of our newsletter

OPENAI A IS BEING ATTACKED ON ALL SIDES IN THE SAFETY DEBATE

Italy, Australia, Germany, Hong Kong, Twitter trolls, Reddit users, a partridge, and a pear tree are part of a growing list of ChatGPT critics raising alarms and calling for more safety in AI development.

OpenAI thinks these people are making mountains out of molehills.

At least, the new blog post on their approach to safety sets that tone. The short read outlines the numerous ways the researchers behind these AI models work to reel in the risks.

What were the main points? The post covered the most prominent concerns and lobs that critics threw against the wall, like:

1/ Child Safety 

Their guidelines are already clear on who can signup for chatGPT, and they’re always working to flag inappropriate content that’s not for general audiences.

2/ Data Privacy 

OpenAI is not Facebook. They don’t need your data for ads, and they don’t collect private information to train these models. Or at least, they strive to take that content out of the training material.

3/ Safeguards 

DAN sucked for them from a PR standpoint, but they red team every model and constantly iterate on the protection to adjust for misuse. You can’t plan against real-world use without putting something in the real world.

4/ Factual Accuracy 

This is an education issue. OpenAI admits the model is prone to hallucinations, but they tell you that from the start. People misunderstanding how these models work is the bigger problem in their eyes. 

Safety talk is not going away anytime soon.

Sam Altman and Co. will need to fight this fight in the media. Too many reports spread misnomers and sensationalized headlines.

Keep your head on a swivel and always come to your own conclusions.

SPONSORED MOMENT

Say Goodbye to Zoom Fatigue

Experience a genuine human connection with Ronday - the new platform for collaboration that creates a shared online space unlike any other.

With video, spatial audio, and warm 3D environments, Ronday fosters a sense of presence that's truly unique.

Get Ronday Free!

GOOGLE WILL ADD CHAT AI TO THEIR SEARCH ENGINE, EVENTUALLY

Sundar Pichai is on a media tour, discussing all the AI plans Google has in store.

His latest visit was to the Wall Street Journal, where he mentioned that a tool like Bard would launch natively in the search engine. However, his quote was vaguer than a Terrance Malick movie (look it up, kids). 

He explicitly guaranteed Large Language Model engagement in the context of search. You can see where that wording lets Pichai get away with a lot without promising much.

Two things stick out from the full interview.

When pushed on the possible LLM integration, the Google exec evaded a concrete answer on if a waitlist would come first. In other words, prepare for more “trusted tester” releases before the general public gets their hand on this.

The cost of the AI model's computational power is a big reason for that. If everyone and their mother used it, Google would bare the cost.

At their scale, it’s not feasible to release this to everyone simultaneously.

Look at this graph from the WSJ piece highlighting how many visits Google gets in a month.

Imagine all those people asking Bard questions all the time. If you thought ChatGPT had surge issues, you haven’t seen anything yet.

Google is massive, and any AI tool they release to the public will immediately boost AI adoption. They know this, and that comes with a lot more risk.

The researchers behind all the fantastic AI that Google’s put out know there’s a ton at stake. That’s why the whole team at Alphabet appears to be picking the tortoise approach to foil OpenAI’s blazing-fast hare.

Neither is wrong. The two timelines are simply different.

Though we can all agree Google’s testing approach is a lot more frustrating.

Quick Nuggets

🥤 Pepsi Co. is one of the many big companies embracing AI in its processes

🎥 The tipping point for AI-generated videos is coming sooner than we expected

💼 Hiring? Add AI to your process by following this guide

✏️ School guidelines for AI are already being implemented in Japan

🔊 Adthos raised a round of funding to use AI to finish audio ads

🌍 Global oversight of AI is one possible way to ensure safety in the development of these models

🤜 A small push is all it takes to make machine learning algorithms go awry

🎙️ Voice actors are facing a real AI automation crisis right now

🦠 Panicking about AI might distract us from other present problems like current bio risks

📌 OpenAI President, Greg Brockman, is profiled by The Information’s Jon Victor

🏙️ ‘Manhattan Project’: a breakdown of why so many AI talking heads compare it to the atomic bomb

🤖 Mark Zuckerberg is pulling all his focus at Meta onto AI

🔨 Microsoft is adding Copilot AI to OneNote

📼 Sam Altman sat down with Reid Hoffman for a catch-up on the latest in AI

✍️ ChatGPT’s poetry is weak sauce, according to actual poets

🚫 Deepfakes: a rundown of all the current tricks used to mislead you

🌲 14 charts from Stanford’s AI report help explain the trends visually

🔥 Fresh Products

• Claid AI - create a variety of quality product images (link)

• WhimsyWorks - your kid’s custom choose-your-own-adventure story (link)

• Amy Reviews - summarizes all the reviews on a product for you (link)

• Any Summary - you upload any file and AI summarizes it (link)

• StorySeed - wants to help create stories that actually standout (link)

• AI Buddy - is a WordPress plugin to quickly generates content (link)

• Great Headlines - promises to help you write copy like Don Draper (link)

• Ai Intern - is a Slack bot to help with all the intern duties (link)

• augment - is a personal assistant that anticipates your needs (link)

• Opus - is a community-driven text-to-video platform (link)

• Habit Driven - is determined to help you reach your goals (link)

• Prompt Storm - access and share a variety of ChatGPT prompts (link)

• PixelBrix - enterprised-focused chrome extension to boost productivity (link)

• Big Tech Company - a satirical take on all the sites popping up, fun for your Friday (link)

Good Content, Vanilla Neo

I’m not going to lie, I have no idea why I shared this, but it was too weird to not include in the newsletter 😂

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

Check Out Our Sunday Edition

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

We’ve written about:

• Here Is The Work AI Will Replace Soon

• Get to Know ChatGPT Through Interactive Games

• The Difficulties in AI Alignment

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

Claim the offer below to support our writing 🙂