Everyday AI Podcast – An AI and ChatGPT Podcast - EP 547: OpenAI and Google team up, Meta’s AI powerplay, Microsoft's Vision and more AI News That Matters

Episode Date: June 16, 2025

Meta is going straight superintelligent? OpenAI and Google are teaming up? And Microsoft might kill tutorials? The big AI dogs are at it again, making some head-turning moves in the AI space. Don't get left behind. Join us on Mondays for our AI News That Matters segment. Newsletter: Sign up for our free daily newsletterMore on this Episode: Episode PageJoin the discussion: Have a question? Join the convo here.Upcoming Episodes: Check out the upcoming Everyday AI Livestream lineupWebsite: YourEverydayAI.comEmail The Show: [e-mail] with Jordan on LinkedInTopics Covered in This Episode:OpenAI & Google Cloud PartnershipMeta's $14B AI Superintelligence BetDisney & Universal vs. Midjourney LawsuitApple's AI Translation Across DevicesOpenAI Expands GPT Projects & FeaturesMicrosoft Updates Copilot Vision ToolsAmazon's AI Video Tool for SellersMeta AI App Privacy ControversyTimestamps:00:00 AI Insights: This Week's Headlines04:28 ChatGPT Project Features Update08:42 AI Wednesdays: Updates and Insights11:57 Meta's $14B Superintelligence Investment15:24 Meta's Acquisition of Scale AI17:31 OpenAI Partners with Google Cloud22:41 Apple Translation Integration Delayed26:00 Third-Party AI to Improve Siri30:43 Microsoft Copilot's Quiet Feature Updates33:40 Amazon Enhances AI for Ads37:34 Copyright Lawsuit: Midjourney vs. Studios41:43 OpenAI O3 Pro: Accuracy Over Speed45:54 AI News & Events Week PreviewKeywords:OpenAI, Google Cloud, Meta, superintelligence, Scale AI, $14 billion investment, artificial superintelligence, Midjourney, Disney lawsuit, Universal lawsuit, AI copyright issues, Microsoft Copilot Vision, Copilot AI updates, real-time AI assistance, AI-powered video generator, Amazon video tools, chat GPT projects, GPTs, AI compute, Tensor Processing Units, regulatory scrutiny, AI market competition, OpenAI revenue, O3 Pro model, ChatGPT Pro, AI reasoning models, voice mode, Apple WWDC, AI translation, Edge AI, cloud infrastructure, Meta AI assistant, Apple AI struggles, Siri upgrade, AI legal battle, OpenAI and Google partnership, Meta's AI investment, $10 billion annual revenue, personal data privacy, Meta privacy concerns, AI model updates.Send Everyday AI and Jordan a text message. (We can't reply back unless you leave contact info)

This is the Everyday AI Show, the everyday podcast where we simplify AI and bring its power to your fingertips. Listen daily for practical advice to boost your career, business, and everyday life. Meet Firefly AI Assistant, now live in Adobe Firefly, the All In One Creative AI Studio. Just describe what you want to create and the assistant handles the rest, orchestrating multi-step workflows across Photoshop, Premiere Express, and more in one conversational interface. You direct the outcome. The assistant accelerates execution. Open A.

I and Google are teaming up. Is meta skipping AGI and going straight into superintelligence? Well, they have a $14 billion bet that they're going to do that. And is Microsoft going to kill off tutorials as we know them with their new and improved AI vision capabilities? A lot of questions floating around. round this week, at least when it comes to AI news updates. And don't worry, if you're spending hours a week trying to figure out what the heck is going on in AI and you can't keep up,

that's what we're here for. Welcome to Everyday AI. What's going on, y'all? My name's Jordan Wilson and I'm the host of this thing. It's your daily live stream podcast and free daily newsletter, helping everyday business leaders such as yourself, not just keep up with AI, but how we can use all this information to get ahead to grow our companies and our business. careers. If that's what you're trying to do, you're in the right place. Welcome. This is unedited, unscripted. We do this every single weekday. So it starts here with learning, but how you can actually leverage this to become the smartest person in AI at your company or your department is by going to our website at your everyday AI.com. So there, not only can you and should you sign up for our

free daily newsletter where we recap each day's episode. So if you miss a couple of things today, Don't worry. It's going to be in our free dealing newsletter. But also on our website, there's more than 440 back. Or sorry, 540. I can't even keep track. More than 540 back episodes of everyday AI. You can go watch them all for free. Listen to them all for free.

Read about them all for free on our website. It is like a free generative AI university taught by the world's leading experts. All right. So I'm excited to jump into today's show and talk about the AI news that matters. Good morning live stream audience. Yeah, podcast people. You can always check out the show notes for a link. But we do this.

It's a live stream. So thanks for joining us. Douglas, joining us from Atlanta. Muhammad saying, good morning. Good morning to you. Jose, joining us from Santiago. The Chile crew is in the house. Kyle joining us from the YouTube machine, Joe, and St.

Fort Lauderdale, Florida. Big Bogey joining us on the YouTube. So thanks for joining, guys. Let me know. If you have any questions, As you go along, if you have any hot takes, you know, I know it's a news day. We'll have our hot take Tuesday tomorrow, but love hearing from our live stream crew. Great to be back, Rolando as well.

All right. Good morning, Cecilia, joining from Chicago, like myself. All right, let's get into the AI news that matters this week as I sip on my coffee. Let's get after it. So Open AI has launched a major update to its chat GPT projects feature, making it more power. tool for ongoing research, writing, and planning. So the new update to projects introduces new tools that let users organize related chats, files, and instructions into a focused workspace moving beyond single session conversations. So the big thing now is users can now conduct multi-step deep research from within a project. So it'll only pull information from that project and it will blend instructions, chat history,

within those projects, file and web research for deeper and more contextual work. This one's a big one, y'all. I can't underestimate enough how big this update is. And not only that, but voice mode has finally been added to projects. So now you can have that hands-free brainstorming in document review just by tapping on the microphone icon in any chat inside a project. So plus in pro subscriber. So this is a paid chat GPT member benefit that's rolling out, can have the enhanced memory

and chat as chatGBT can reference past chats within a project to maintain continuity and context. So yeah, if you want your project files to be able to automatically understand the context from other chats underneath that project, that is a paid only feature right now. The mobile app supports file uploads and model switching now, which is great bringing some of the functionality we've had on the web for a while. Also, individual chats within a project can be shared via unique links so collaborators can see just the intended conversation and not the entire project. So OpenAI says all users can create projects now, but only, but these premium features like

memory across chats are reserved for the paid tiers. So pretty big update here. I think a lot of people have had the misconception that projects, both within chat, GBT and Anthropics Clawed, are more just for organizing your chats, right? Oh, they're a folder structure like a folder on your operating system for your computer. And yeah, they kind of are, there's so much more than that, especially after this new update from Open AI. I think probably the two biggest ones just within this project update is number one, being

able to interact with your voice now, which is really big. But the biggest one would probably be the deep research just within the project, which is huge, right? So now you can dump a ton of your company's information files documents. And now it searches across past chats as well for paid users. So pretty big one here. And I'm curious if any of our live stream audience is going to be or has already experimented with it. I have definitely. Renee says, thank goodness, I need this for a project.

I think this is, if I'm being honest, just a pretty simple, I won't say simple. I think this is a fairly straightforward functionality that probably should have been there from the beginning, but I'm sure there's some technical issues to it. And I would assume that all the other big players are going to have to roll out this functionality, because like I said, this just greatly improves the way that these, folders can work. So not only that, also a big update that Open AI didn't even really release, but it's out. And maybe you could argue it's even bigger.

GPTs, yes, OpenAI's custom version of their model called GPTs can now use any model. All right. So what that means is you've been able to create for a long time, I think almost now two years, a version of chat GPT called GPTs, right? So it's a smaller version. You can put your own custom instructions and files in there. But the biggest difference between GPTs and folders is you can mention or call a GPT, no matter where you are as long as you're using a GPT model.

It doesn't work with the O models. But the biggest differences is you can be in any chat and you can reference essentially an unlimited number of GPTs one by one. The downside is they've been on this older model, GPT4O, which is still. a very capable model. But now you can use the O series models when using GPDs, which turns them from, okay,

this is a pretty nice feature into, okay, we have to redesign how we work at our company. I kid you not. It is that big, the ability to use these O models, these thinking models, the ones that can reason and plan ahead

and take time to think about a very complex query with a lot of data and a lot of instructions. So two pretty big updates there from OpenAI. Speaking of which, Make sure to join us Wednesday. We're doing the new series, I don't know, do you guys like it or not? Putting AI to work on Wednesdays. So we have our news Mondays, our Hot Take Tuesdays, and we're going to put AI to work on Wednesdays.

So this Wednesday, we're going to talk about this from OpenAI as well as comparing some of these new updates across Google's, Gems, chat GPTs, GPs, and the projects from Open AI and Claude. So we're going to be doing custom GPs, gems, and products. projects, what they are and why you need to use them. All right. So make sure to join us on Wednesday for that show. All right. Our next piece of AI news, Meta's new AI assistant app, which launched in April, has

already passed one billion users according to Meta's own announcement. And despite Meta's claim that nothing is shared to your feed unless you choose to post it, the app's interface makes it easy for users to. accidentally share private queries publicly via a Discover feed. And this thing has been going viral. Essentially, got a bunch of people that don't understand that if you share something in a certain way, it becomes public.

So essentially, you have thousands and maybe more of users who are thinking they're chatting with Meta's AI assistant, but they're actually posting these things publicly. So a review of the Discover feed by Futurism in this case, which is the news article we are showing here for our live stream audience reveals that highly personal and sensitive questions, including medical issues, sexual health concerns, and legal troubles are visible to the public, often under users real names if they logged in with Facebook or Instagram.

Yeah, not good. The app also uploads and shares lengthy voice queries without clear warning, with some recordings lasting over an hour and containing personal and controversial discussions. So examples found in the public feed include users asking about bowel movements, recurring STIs, tax problems, and even advice for criminal sentencing letters, all with identifying information attached. Also, teachers, students, and professionals are unwittingly sharing work-related queries

such as test questions and advice for handling, relationships, making the app a potential privacy minefield. So yeah, you technically have to consent to posting this publicly, but it's not 100% clear, especially if you're not a super technical user. So yeah, the new meta AI app is actually turning into, I guess, must see or must read social media, but unfortunately, due to people not fully understanding and thinking their queries with this AI chatbot are private and their public. This is not going to turn out well for meta.

It's going to become a PR disaster because this just started kind of some of these details just kind of started leaking out some sleuths online are like, wait, is anyone else seeing this? And then everyone else is like, yes, we are. Some of the media started to pick up on it this weekend. I would expect the mainstream media. had to pick off on this over the next week or two, and it's going to turn into a huge story and a disaster for meta. All right. Speaking of meta, they're making a $14 billion investment into

not just scale AI, but super intelligence. So meta has finalized a 14 plus billion dollar investment into scale AI, valuing the data labeling and model evaluation startup at 20, billion, according to reports from the New York Times. So this marks Meta's second largest investment ever after its $19 billion Wats app acquisition. So as part of the deal, scale AI. So again, if you don't know, scale AI, there's a good chance you don't unless you are a geek and like me and father the AI scene, but they are probably the biggest data labeling

company that so many of the AI labs use to help them clean up and improve their data and they also do model evaluation. So as part of the deal, scale AI, scale AI founder and CEO Alexander Wang will join meta to lead its new super intelligence team, a team focus on developing AI with capabilities far beyond current systems. So if you're wondering like super intelligence, like what the heck is that? Well, uh, essentially you have artificial intelligence, which is the stage that we're at now, uh, right? You have artificial general intelligence or AGI, which is essentially when the AI systems that we all use are as good at everyday tasks that humans are. And I would argue that we've kind of been there unofficially for the last year or so.

Depends on what your definition of AGI is. But ASI or artificial super intelligence is what happens when the machines are far more intelligent than human. So it kind of goes AI, AGI, ASI. So meta here is placing a huge. bet into creating this $14 billion kind of acquisition of scale AI or aqua higher, you can say, as a lot of scale AI leaders are now going to be joining this new meta super intelligence team. So several scale AI employees will move with Wang to meta, where they will work directly on advanced AI projects.

So meta will own a 49% stake in scale AIs continued operations, but will reportedly have limited control over its operations, a structure designed to avoid regulatory scrutiny that has affected similar deals involving Amazon, Microsoft, Google, and just about everyone else. So the New York Times reports that meta CEO, Mark Zuckerberg, has been personally recruiting top AI talent, even inviting candidates to his home

and offering compensation packages worth up to nine figures. Yeah, I'll do the math for you. That's compensation packages at valued at hundreds of millions of dollars. So all you, you know, people that maybe got your computer science and AI degrees like 20 years ago and people are like, oh, what's this all about? Are you sure there's a job in this? Yeah, now you guys are making salaries that make NBA salaries look minuscule. All right. So Wang described his move as a significant opportunity, though it comes at the cost of leading.

his CEO role at Skill AI, but Jason Droge, the startup's chief strategy officer, will take over as CEO. Wang will remain on Scale AI's board of directors, ensuring ongoing development with the company. So this is going to be pretty interesting how this shakes out, because this isn't just meta, making a huge investment into technically AGI, ASI, and beyond. it's actually probably going to impact all of the other AI systems that we use. As an example, recent reports are showing that Google will reportedly cut ties with scale AI and then we'll have to see what's open AI, Google and Nvidia and others are going to do with their partnership, right?

Because just about every single big AI company and every single big tech company has partnered with scale AI. So now that Scale AI has essentially been Aqua hired by Meta, you've already seen Google reportedly is going to be cutting their ties with Scale AI. So we'll see what OpenAI, Microsoft, Nvidia, and others do in terms of their partnership. Because even though Meta is saying, like, yeah, we're going to kind of keep these things separate, I don't know. If Wang is on the board of directors of Scale AI, but is leading the super intelligence teams,

I don't know. I would assume that some of the other big companies will follow Google's lead in cutting ties with Scale AI. And I would assume that Scale AI's closest competitors are going to be gobbling up new business opportunities because, yeah, you're going to have trillion dollar companies being like, yo, we need another partner in this space like yesterday. Who can get us up to speed? So pretty big movements there from the big players. Speaking of movements from the big players, yeah, Open AI and Google are partnering together. What? These are the two biggest companies going head to head, punch for punch.

I'd say since December of last year, it's been a battle who is the AI king. But now they are shaking hands, at least when it comes to cloud compute. So Open AI has finalized a deal to use Google Cloud to expand. and its computing capacity, according to reports, marking a significant shift in artificial intelligence industry's competitive landscape. So yeah, this is according to a Reuters report. So this move allows OpenAI to reduce its reliance on Microsoft Azure,

which has been its exclusive infrastructure partner until earlier this year. So yeah, it looks like Open AI is open to working with other tech giants, even though Microsoft has a reported 49% stake in OpenAI after reportedly investing up to $13 billion. So also announced as part of this, Open AIs annualized revenue run rate has surged to $10 billion as of June, highlighting the rapid adoption in growing demands of the technology. So this new partnership between Open AI and Google is seen as a big win for Google Cloud, as it will supply additional computing power to Open AI for training and running AI models, despite OpenAI's chat CBT posing a direct threat to Google's search business.

Right. So yeah, this on certain aspects, this makes a lot of sense. And on certain aspects, it doesn't, right? Because, yes, this is going to be bringing in a ton of revenue to Google Cloud. But also, they're a direct threat to not just Google's search business. Open AI is a direct threat, not just Google search business, but to their Google Gemini product. So interesting that we're seeing this kind of partnership unfold. So this deal comes as Google expands access to its in-house TPU or tensor processing units to external customers,

including other high-profile tech companies and AI startups. So Open AI's recent partnerships, including the $500 billion, Stargate, infrastructure project with SoftBank and Oracle and multi-billion-dollar deals with CoreWeave show its ongoing efforts to diversify its computing resources. Both Open AI and Google declined to comment on the deal, but Alphabet's stock rose 2% after the news, while Microsoft shares dipped by about 0.6%. Analysts at Scotia Bank called the development somewhat surprising and noted. It's demonstrated how fierce competitors are willing to collaborate to meet massive infrastructure needs. So yeah, pretty shocking, I'll say, when I saw that headline early on the week and I was like, huh, interesting, right? I guess it does show that number one, money talks and compute is like the biggest resource right now.

right so obviously open a i making ways on the revenue side uh google with their tp u's a lot of their announcements that they just announced at their google cloud uh next and just their their tpus and just the uh how they're handling and expanding and scaling uh cloud compute pretty impressive so i guess it makes sense but hey competition is at least got to take a break in this uh in this regard all right uh let's take a quick break for me to sit my coffee and hear from our partners. This podcast is supported by Google. Hey everyone, David here, one of the product leads for Google Gemini.

Check out VO3 or state-of-the-art AI video generation model in the Gemini app, which lets you create high-quality, eight-second videos with native audio generation. Try it with a Google AI pro plan or get the highest access with the ultra plan. Sign up at Gemini.com Google to get started and show us what you create. All right. I like this comment here from Joe saying cloud infrastructure diversity is good for Open AI. And I suspect it is a trend to snap up more computing resources. Compute resources. Yeah. Great observation there. Joe. All right. Our next piece of AI news, while all the other companies are making big AI splashes at their yearly conferences,

Apple obviously came out of their WWVC, their worldwide developer conference last week with a huge. AI yawn. But one thing that does look pretty cool and pretty promising at least is Apple has unveiled their forthcoming live AI translation across devices. So at WWDC, Apple introduced the live AI power translation as a central feature across its platforms powered by the new Apple intelligence engine. So translation is, will be integrated. on supporting devices across messages, FaceTime, and phone, enabling real-time text and speech translation in up to 17 languages, though supports varies by app and mode. So live text translation in messages displays the translation instantly as users type, while

FaceTime offers speech-to-text translation with live captions during video calls. On phone calls, users can access both speech-to-speech and speech-to-text translation, with a synthetic voice reading the translated text allowed in a different language as it appears on screen. At launch, it languages are supported in messages in five in phone and FaceTime, with Apple promising to expand to eight more languages by years end. Reports, though, are showing that the new and smarter Siri that we've all been hoping for and waiting for is going to be delayed until 2026, while Apple, I guess, continues to struggle with it and handle all the class action lawsuits from promoting all these smart AI features last year and rolling out

pretty much none of them unless you're excited by using AI to create emojis. And then, oh, this year, the big update there is you can use AI to combine two different emojis. Right. Gosh, I still don't understand how Apple has fumbled the AI bag so hard. But I can't wait until someone makes a movie about this and how this is probably going to be one of the reasons why Apple has released its stranglehold on being the biggest company in the world. I personally don't see Apple ever reclaiming that number one spot. And I would be surprised if in three years they're even a top five company in the world anymore by market cap.

So I would say that its inability to capitalize on AI has led to its tremendous downfall, which I detailed last week. You know, if you want the hot take, I think last week's was a good one. I know it was about a research paper, but go listen to episode 543. The title of that one is Apple's weaponized research inside its illusion of thinking paper.

Yeah, essentially Apple researchers came out with this quote unquote research paper, which was actually just marketing because it came out like two days before their WWDC conference where they essentially took a gap year on AI. Right. So they came out of this paper that essentially questioned like, oh, these large language models, that reason, they don't actually reason.

I ripped it apart. It was a terrible, terrible study. And actually, there's been follow-up studies by other companies saying that this is a terrible study, including one which I liked called the illusion of the illusion of thinking. So that one was pretty fun. But one or two other small AI announcements from Apple from their WWDC conference. They did say they're opening up their core Apple. intelligence AI models to third-party Apple developers, allowing any app to tap directly into the

on-device language model. So, yeah, if you have a newer Apple iPhone or a newer Mac or MacBook, you will have access to their kind of edge AI or on-device large language models. So now developers will be able to access those on-device AI models without really having to have the AI capacity themselves. So what this could mean is, in theory, we could. actually get a smart Siri, but from a third-party company, which I would laugh. And I would use it, obviously, because Siri is useless.

It still can't even tell the time or tell me directions or tell me what the temperature is because it doesn't understand anything and there's no ability to talk to it. So I would hope that some company out there, and I know perplexity has already started to kind of carve out that niche. So please, someone out there, go build a smart Siri that, you know, a billion or two billion, however many Apple devices there are out there can actually start using. Yeah, another good comment here.

Apple translation jokingly reminds me of Monty Python's Hungarian phrasebook sketch. All right. All right. Some Microsoft news. And maybe this could kill off tutorials. Maybe at least in some instances, I think it actually might or just this movement of AI technology.

But Microsoft has announced the release and technically update of co-pilot vision on Windows with highlights. Now available in the US for Windows 10 and 11 users, according to Microsoft's official announcement. So this update is notable because co-pilot vision lets users share what's on their screens with the AI assistant, which can then provide real-time insights, guidance, and answers based on the content they're viewing. So this is if you've used Google AI Studio like stream real time, very much similar to that.

You can essentially share any part of your screen with Copilot Vision. It can see, it can talk to you in real time. You can ask questions and it's contextually aware with what's on your screen. So Copilot Vision can analyze content from up to two different apps at once, which that's big, helping users connect information and navigate tasks more efficiently across multiple Windows. So the new highlights feature, which I think is pretty cool, allows users to ask co-pilot to show me how for specific tasks with step by step in app guidance, making it easier to learn new workflows or troubleshoot issues. So that's essentially it will highlight things on your screen.

So as an example, let's say you're trying to learn something in, I don't know, Adobe Photoshop. And you're like, hey, how do I mask this layer? Right. You can just open up micro, Microsoft's co-pilot, launch co-pilot vision, share the screens that you need to and talk to it. And it will go and literally walk you through step by step. It will highlight things on your screen. It'll say, you know, go into this menu, hover over this, click this, you know, this keyboard shortcut. So literally, I like, I think there are so many, you know, I guess online entities, you know, YouTubers that really just exploded over the past decade doing just this, right?

probably even longer than a decade, just doing software tutorials. And I don't know. And from being honest, I think the whole tutorial landscape might kind of shrink. Because as long as players like Microsoft, Google, and whenever Open AI actually announces this, even though they teased it more than a year ago, I think it was April or May of 2024, that they teased this. As long as these three companies specifically continue, to invest in this space, I would assume this is how the majority of us business leaders are going

to be interacting and learning new softwares, new processes, et cetera, right? But that also requires that you really understand these models, what's their knowledge cut off, how do they connect to the internet, how can they bring in your own personal context to help you with these tasks? So the update also introduces deep research and file search capabilities within the co-pilot app expanding its usefulness for both work and personal projects. Also, Microsoft plans to expand co-pilot vision availability to more countries outside Europe in the near future, potentially broadening its impact in adoption. All right.

Hey, big bogey said go, co-pilot. Yeah, I'm excited because here's like what I've seen happening. Microsoft has just quietly been. updating its kind of AI products after, you know, Open AI in Google. So there's actually a ton of capability that has quietly been rolling out across Microsoft co-pilot's products that I don't think they're necessarily getting a lot of love for. And maybe that's because they're, you know, other companies are doing it first, right? So, you know, as an example, you know, Google with their notebook LM, their audio overviews,

these deep dive AI podcasts, right, where you could dump in all your context, click a button, and these two AI hosts will create a personalized podcast just based off your content. Microsoft rolled out that feature as well like a month ago, right? So Microsoft has quietly been updating its co-pilot to bring in some of these other capabilities. So it's been interesting. Essentially, you know, you have Open AI and Google going toe to toe.

And then Microsoft either weeks or months later, just kind of borrowing. some of the best features and functionality that Google and Open AI are rolling out. So it's been it's been kind of fun. It's been kind of fun to watch. So Brian with a good question here, how is this different than what I can do with Google's AI Studio with screen share? Great question, Brian. So a couple of differences is this works as a desktop app versus when you're using Google's

AI studio with their stream real time with screen share. It's you're working in a browser base. The other difference is with co-pilot vision, you can share multiple apps at once. Whereas with Google's AI Studio, you can only share one screen at a time. Although I haven't tried sharing a single screen that has multiple apps running on it, but this new co-pilot vision feature by default lets you share up to two apps at a time. So let's say, I don't know, you have two different programs open. in maybe an audio editing program

in a video editing program and you're trying to, you know, sync something up between the two. So there's a lot of instances where you might want to be able to share two different apps and not just have one screen sharing. But yeah, well, as this rolls out

and as the technology matures, we'll probably want to do a head-to-head comparison at some point. But great question there, Brian. All right. Our next piece of AI news, this is either great or terrible depending on how you look at it.

But Amazon has launched its AI-powered video generator tool for U.S. sellers inside Amazon after nine months of beta testing. So yeah, we talked about this in late 2024 when it was first reported, but it looks like now this tool is being rolled out. So the free tool for Amazon sellers allows authenticated Amazon seller accounts to create eight seconds, what they call quote unquote, low motion video ads from exactly. existing product images, making video advertising more accessible for small businesses and online Amazon sellers. So Amazon has also updated its AI image generator, which now better places products like wristwatches on photo realistic human models, advancing beyond simple white background product shots. So the new and updated video generator can summarize key clips from existing videos, but also generate new visual ads, helping sellers with limited marketing budgets or expertise.

Half of the products promoted with AI generated video assets are being advertised for the first time, suggesting these tools are enabling new advertisers to participate. Amazon reports that sponsored brand campaigns featuring AI or sorry, featuring video elements see an average of a 30% higher click-through rate, though the company has not disclosed whether, this leads to higher sales overall. The growing use of AI generated video in ads is rapidly expanding across Amazon's catalog of product images and videos further improving its generated AI models.

So while the tools can now produce short montage style clips with human figures, they are still very limited in creating longer form narratives. And also, the quality, not that good, right? If you're thinking you're going to see quality similar to Google Vio or clings or even, you know, Open AIs, SORA, not so much. You know, don't expect any jaw dropping, you know, videos to be dropping anytime soon or if you're an Amazon seller. But what this does mean for consumers is you just have to be cognizant, right? Even when I'm buying something off Amazon, I generally don't look at the photos.

I go straight to the videos, right? That's me. But keep in mind, there's obviously a such thing as video hallucination. Luckily, it seems like Amazon is rolling this out slowly and is focusing on more what they're calling, which I kind of chuckled that they said low motion video ads. But keep in mind, video hallucinations are a thing. So before you go base a decision on something, especially if it's expensive that you're buying on Amazon by watching an AI image or sorry, an AI generated video, make sure that that's actually factual, right? and make sure that whatever capabilities are being described or shown in that AI video are not actually just hallucinations.

All right. Two more pieces of AI news and our two last ones are pretty big. All right. So this could be the first AI lawsuit that starts to really hit the AI companies and that could stick. I'll say that. So Disney and Universal have filed a amazing. lawsuit against AI image generator Mid Journey, accusing it of generating endless unauthorized

copies of their iconic characters. So the studios allege that Mid Journey made $300 million last year, but by a big part of it was by allowing users to create images resembling their own copyrighted characters, such as Elsa, Shrek, and Iron Man, often with striking visual similarity, not quite pixel for pixel, but pretty dang close. And this marks the first time a major Hollywood studio has taken direct action in the ongoing debate over AI and copyright, a move that could set a crucial legal precedent for the entertainment industry. I'm curious.

I put out a poll on this on my LinkedIn, but live stream audience, you've seen this, you've seen mid-jury, you all follow the AI space. who do you think is going to win this? Do you think that Hollywood is going to win this case? Or do you think Mid Journey is going to win it? Let me know in the comments, but a little more about the lawsuit. It is a 110 page complaint, and it includes dozens of side-by-side comparisons that show what appeared to be, again, I'm no legal expert, but I did say at a holiday inexpress once. It looks like just, I mean, side-by-side. It's like, yeah, that's a copy, right? But it shows dozens of side. by size comparisons of Mid-Journey outputs and the original

copyrighted images from Universal and Disney. And the studios argue that Mid-Journey could filter out copyrighted content if it wanted to, just as it currently filters out violence and nudity, but has refused to do so, they say, even after they

sent them cease and desist letters. The lawsuit seeks damages or statutory penalties of up to $150,000 per inflation. fringed work as well as an injunction to block midgerty's image and their just released or slowly being rolled out anyway new video services until copyright protection measures are in place. So I don't know if that $150,000 per infringe work would be retroactive.

But if it would, I would assume that would be like trillions of dollars, right? there's been millions of mid-journey images produced, and I would assume the overwhelming, okay, I won't say the overwhelming majority of them, but I will do say there's probably countless of millions of images that have been created against Disney or very similar to Disney and Universal's lawsuit. So that $150,000 per infringed work, I'm curious if that'll be retroactive or only if the ruling is in their favor, if it only applies from there forward. So that one is going to be interesting to watch.

And Mid Journey's founder has previously admitted essentially that the company did not seek permission from copyright holders, citing the impossibility of tracking the source of millions of images that it has scraped from the internet to train its models. The case highlights a broader industry conflict while Hollywood studios are suing to protect their intellectual property. Many, including Disney, are also exploring responsible use cases of AI to cut costs and boost creativity. The outcome could impact not just studios and AI companies, but also artists, writers,

and anyone whose work or lightness might be used by generative AI without consent. This is going to be a big one, y'all. And so big tomorrow, we're going Hot Take Tuesday style on this one. Disney versus Mid Journey will Hollywood kill AI. And I've got some takes that you're probably going to want to listen to, including apparently I'm the only one because I've read all the study, like I've read all the news reports on this. I've been tracking Mid Journey's terms of service since I started everyday AI.

I've done a show on this. But essentially more than a year ago, they changed their terms of service. And guess what? They, well, are trying to wash their hands. because they said, hey, if any company ever sues us, we aren't responsible. You are responsible. The user. It literally says that plain as they in their terms of service.

So we're going to be breaking that and a lot more down on this story tomorrow. It's going to be an interesting one to see how this plays out because this might be one of the first times that I'm looking at this and being like, all right, this one looks kind of cut and dry against the AI company. When we're talking about text-based large language models and lawsuits they're facing, I think it's much harder to prove. and the gray area is enormous. I don't know. This one looks a little cut and dry to me. Again, I'm not a legal expert.

All right. Last but not least for our pieces of AI news, OpenAI has released the much anticipated O3 Pro model for Chad GPT Pro and team users with Enterprise and EDU customers gaining access next week. So the O3 Pro model is designed to provide more reliable and accurate answers by spending additional time on challenging queries, offering improved performance over both the standard 03 reasoning model and the 01 Pro model.

So OpenAI recommends using 03 Pro for situations where answer quality is more important than speed as responses could take a few minutes. Even simple responses. Don't accidentally say like, what's up to 03 Pro because it might take like five minutes to really ponder what is actually up. But benchmark testing shared by OpenAI shows that O3 Pro outperforms previous versions by a lot, making it a strong option for complex, a strong option for users who need dependable results for complex or sensitive questions. Some features, including image generation and canvas, are not supported on the O3 Pro model for now.

So users would need to switch models for those tasks. The rapid rollout of new models, including O3 and O4 Mini in April, open AI's ongoing push to expand their reasoning models and improve the user experience. Open AI also announced a huge 80% price drop on the base O3 model, not O3 Pro, but an 80% price drop on the O3 model in the API, which I think is one of the more consequential moves on the developer side over the past like two years. That piece is actually huge. So we're actually going to be running down in tomorrow's show as we go over

Google gems versus GPTs versus projects from OpenAI and Claude. We're going to be using the O3 Pro model. So make sure you tune into that show to see how the results end up. All right. And since you guys said you liked this, we're going to go over rumors and what's next. All right. So these things, some of these things have been announced. Some of them are rumors.

Some of these are starting to roll out. And we'll probably be talking about them more in the coming weeks as they actually start to hit accounts. So Google is rolling out its plan ahead scheduled actions inside its Gemini chatbot, which I'm extremely excited about. Imagine scheduling ongoing tasks to check your email, calendar, Google Docs, etc. That one I'm super stoked about. Open AI has said their open source model.

They're delaying it because they came up with some huge breakthroughs. So originally reports are saying it might be in June. Now they're saying it's just going to be later in summer. Google Search Lab. has rolled out or started to roll out an audio overview to where you can listen to an AI generated audio overview on certain search results. Perplexity tasks are starting to roll out to paid users. So, yeah, a lot with this scheduling certain tasks or scheduling certain prompts in large language model, that's been a huge wave over this second quarter here of 2025.

ChatGPT has upgraded and we've seen this starting to roll out its base search functionality. And in some cases, we're seeing a slight chain of thought or reasoning in its non-reasoning model, GPT40. So that one's interesting. A co-pilot recorder is starting to roll out to beta testers, which, you know, we talked about the copilot vision. Essentially, you can record certain tasks, and it will go out and then do them for you on your behalf, which is similar to a feature that Google's Project Mariner has, which, my gosh, we got to do a show specifically on that one. because I think it's really cool. Grock is slowly rolling out file images and also scheduled tasks support in its sidebar.

And Google is reportedly rolling out new models this week. This is crazy, y'all. So we might see a Gemini Pro on, I forgot which one's Tuesday and which one's Thursday, but we might be seeing a Gemini Pro and a Gemini Flashlight. So two new updates to Google's most powerful, 2.5 Gemini models could be rolling out this week. My gosh, a ton, a ton to go over. All right.

So like I said, make sure to join us tomorrow for Hot Take Tuesday on the AI Studios versus AI mid-jurney and AI studios in their big legal fight. And then join us on Wednesday for putting AI to work Wednesday as we look at Google Jems versus GPs. versus custom projects in Claude and chat GPT, but a very quick recap on the AI news that matters for this week. So number one, OpenAI has expanded its Chad GPT projects and GPTs. Meta's AI app is causing a little bit of controversy

as users are seemingly accidentally posting their private queries with the AI chatbot to be publicly discoverable. Meta has invested $14 billion in both scale AI. and its founder as they start their new super intelligence lab. Open AI and Google Cloud have struck a surprising but also seemingly sensible deal to team up on AI compute. Apple announced their WWDC AI announcements, nothing big but live translation and allowing developers to use their audit device edge models to have better apps. Microsoft has updated and launched its co-pilot vision on Windows, offering real-time AI assistance. Amazon has expanded its AI video tools to create AI videos for online sellers.

Disney and Universal are suing Mid Journey. Tune in tomorrow for that show. And last but not least, OpenAI has launched their O3 Pro model for paid users. My gosh, this was a ton. If this was helpful, please go to your everyday. AAI.com, sign up for the free daily newsletter, but also, please go click that repost button. And maybe this is the one time of week you tune in. If so, thank you. Like I said, we're trying to provide a little bit more structure on our weekly lineup. So we have our Monday, AI News

That Matters, Hot Take Tuesdays, putting AI to work Wednesdays. And then Thursday and Fridays, we're going to bring on amazing guests for you all to ask questions and just to learn from. So if you're listening on the podcast, thank you for tuning in. Please subscribe and follow. the show, do me a favor and do that. But also, if you are on the live stream, please click that repost button and share this with your network. So thank you for tuning in. Hope to see you tomorrow and every day for more everyday AI.

Thanks, y'all. Meet Firefly AI assistant. Now live in Adobe Firefly, the Allman One Creative AI Studio. Just describe what you want to create in your own words and the assistant handles the rest, orchestrating multi-step workflows across Adobe Creative Cloud apps, including Photoshop, Premier Express, and more. one conversational interface. You direct the outcome while the assistant accelerates execution.

Stand control with the ability to step in and refine at any time. See it today at firefly.adobie.com. And that's a wrap for today's edition of Everyday AI. Thanks for joining us. If you enjoyed this episode, please subscribe and leave us a rating. It helps keep us going. For a little more AI magic, visit Your EverydayAI.com and sign up to our daily newsletter so you don't get left behind.

Go break some barriers and we'll see you next time.