Issue #18: Marketing Hollywood with AI, ChatGPT gets live search data, fake news, AI tutorials, and the future of SEO
The rise of AI-generated images hit the headlines this week after the US stock market took a dip in one of the highest-profile cases of fake news. The Times of London - among many other media outlets - reported the spectacular reaction to the AI-generated image.
People were thrown by the image showing a big plume of black smoke near a building looking a little like the Pentagon. It appeared on a “verified” Twitter profile, but Elon Musk’s new system means the authenticity checks are skipped in favor of a paid verification model.
“Large Explosion near The Pentagon Complex in Washington, D.C. - Initial Report,” read the caption.
The initial misunderstanding created by the image was debunked by authorities within about 20 minutes. And the stock market recovered soon after.
But the damage was already done when it comes to generative AI being labelled as the source of disinformation.
While this meant the incident captured the attention of the media, it’s worth remembering that demonizing the tool is like, well, I’ll just leave this here:
In today’s issue:
AI goes to Hollywood: How the robots behind the camera are getting eyes on the ones in front
ChatGPT gets access to current search data
AI art tutorials to make marketing asset creation a cinch
How to transform your blogs into videos in 2 minutes
SEO is dead. Again?
Let’s dive in.
Google is dropping AI bombs everywhere (I gave you a taster last week), and this week the CEO is staking the company’s claim as an AI OG, reminding us they’ve been incorporating AI into their products for a decade.
Alphabet and Google CEO, Sundar Pichai is also keen to emphasize Google’s focus on AI responsibility: “...what matters even more is the race to build AI responsibly and make sure that as a society we get it right.”
He said, “AI is the most profound technology humanity is working on today”.
And Google’s launch of the next-gen large language model PaLM 2 does demonstrate machine learning’s capabilities on a totally unprecedented scale.
Right when it launched, Tech Crunch released a paper that explained the intelligence behind the LLM’s dataset. And it’s bad news for Microsoft because it goes way above and beyond that of even the latest version of ChatGPT.
Although the co-authors kept their cards close to their chest, they did reveal that the bot was trained using a collection of websites, research papers, conversational data, and code repositories, in combination with an extensive resource of non-English text.
But what most don’t know about Google is…
In 2018, the tech giant’s Advanced Solutions Lab worked with 20th Century Fox's data science team to develop market segmentation and map out viewer segmentation.
This allows them to grow their understanding of what viewers are looking for, giving them more predicted film viewing patterns for future film launches. With this access to machine learning paired with the more robust deep learning, they engineered a way to particulate viewer data and scripts.
In 1.5 years, they could ID basic patterns that are now standard for creating movie tone that is “data-driven” to engage higher ROI on movie budgets.
As the article above shares, this is how production companies have been able to scale leaps and bounds over the past decade: “Understanding the market segmentation of the movie-going public is a core function of movie studios.”
Warner Brothers signed a deal with Cinelytic in 2020 to catch up… and just in time. “The integrated online platform can assess the value of a star in any territory and how much a film is expected to make in theaters and on other ancillary streams.”
If you want to become more familiar with machine learning and the difference between it, AI, and deep learning so you can grasp how layering them was the catalyst for how we view movies (and how you can leverage it for your VOC), check out this deep dive by RingCentral…
Wanna make your own visuals and videos, amp up your content and video creation so you can slay at Youtube, Meta and Google Ad creatives?
Thanks to the wide range of online AI art tutorials, you can learn how to use algorithms and neural networks to create unique art pieces for anything (even turning them into short video ads).
This step-by-step guide by Creative Bloq offers detailed instructions and examples for Midjouney, DALL-E 2 and Stable Diffusion and more.
It’s a great one if you want to start using AI for ads, realistic portraits of products - anything you can dream up, really.
It doesn’t matter if you’re a designer or a traditional artist looking to incorporate AI into your “mixed media” mediums, this tutorial offers a great starting point.
And if you can’t get enough of AI tutorials, games, and learning resources, then check out Google’s Machine Learning suite. You can also discover how to create your own customized large language models, create unique, generative concepts, and compute intelligent data graphs for analytics and marketing purposes.
Between this and AutoGPTs, products and marketing are about to transform - and much faster than we have ever seen before.
AI is like Gizmo after a late-night snack… little AI’s popping out, disrupting every corner of the market.
We now have a solid idea of how Hollywood uses AI to streamline their production procs and amplify their marketing techniques.
And one thing Hollywood is good for is clickbait headers and drama…
But the gossip about the Midjourney updates - MJ 5.1 - is spreading like wildfire and people are taking to the streets (get to that in a minute).
The latest iteration of the generative art AI service comes with a range of features, including a raw, “unopinionated” mode designed for more realistic images.
It’s safe to say the improvement between the previous iteration and 5.1 is significant.
The realism, level of detail, and most of all, credibility have all taken a big leap forward.
Far from needing long, complicated prompts, Midjourney 5.1 now responds to simple prompts and:
Produces sharper images
Reduces unwanted borders
Reduces text ”artifacts”
Expert users can dive into using “raw” mode for more creative control.
And now, it’s not just the Pope donning one of those chic Balenciaga jackets…
And speaking of the “Dark Side”...
AI’s disruption made late-night television shows go dark across the network following a mass strike, which has seen over 11,000 television, film and screenwriters taking action over poor job security and diminishing pay.
The studios behind the country’s top late-night talk shows, including The Late Show with Stephen Colbert and Jimmy Kimmel Live are among those impacted.
One of the primary motivations behind the strike is the rise of AI within the film industry.
Specifically, writers are concerned Hollywood executives will overlook AI’s potential as a tool, and instead, use it as a cost-curbing, profit-maximizing replacement.
All most writers want is to integrate AI into the screenwriting process - not necessarily ban the tools altogether.
By demanding protection from AI-driven redundancy and letting the studios use it properly, the talent can continue to grow and learn as they strive to create mind-altering media within the technological revolution.
Then there is this…
Deepcake, the world’s first AI-powered talent agency - specializing in bringing both deceased and retired actors back into the limelight with generative tech.
Their most recent recruit? Bruce Willis. (Influence without the influencer?)
Somehow, what they used the Hollywood legend’s likeness for is more chilling than Sixth Sense: they used their Deepcake Willis for a Russian smartphone ad.
But with politics aside, it’s a fascinating peek into what the future of advertising might resemble. Soon, we could be using Leonard Nimoy to promote Bionic Marketing on a YouTube ad.
But for those who don’t consider themselves “Die-Hard” Trekkies - the same ad could just as easily feature Mark Hamil, Sigourney Weaver, or Brad Pitt…
Because AI has the potential to determine your tastes and what will capture your attention - and then generate a commercial for a given product based on that data.
$2B: Yes - you read that right. Lending platform Upstart has been the recipient of a cool two billion dollars for its AI-powered verify and process loaning system. They claim the system will help businesses navigate economic slowdowns.
$100M: Automated vehicle inspection is now a reality thanks to UVeye. This revolutionary approach to car manufacturing is the first fully automated quality assurance system, and will be made available for automakers, dealership groups and used-car auctions alike.
Does the thought of repurposing that backlog of valuable (and often lengthy) blog posts into video content hurt your brain?
Well, there’s an AI for that… (without it, it’s tedious, expensive and challenging - especially when voice actors, software costs and footage are factored in.)
🤖 With Fliki.ai, you can transform your blog posts into polished, professional-looking videos in just 2 minutes. Best of all - the app’s flawless AI-generated voiceover sounds just like a real person.
No need for any special video editing skills. Simply enter your blog post URL, customize the video settings to your liking, and let AI handle the rest. Whether you're looking to repurpose your existing content or create brand new video content, Fliki makes it easy and accessible for everyone.
So, no more settling for written content when you can burn and churn it into a visually stunning video in minutes.
🤖 If you’ve ever wanted your very own HAL to give you your morning digest, you’re in luck: ChatGPT-powered Inbox Narrator is a powerful tool that automatically turns your emails into audio files.
With AI-powered text-to-speech, you can listen to your emails like they are podcasts - and even respond with your own voice.
The Narrator offers integrated multi-language support in a selection of different voices, so you can finetune the app exactly to your preference. Plus, it's compatible with popular email services like Gmail, Outlook, and more.
🤖 I mentioned Meta’s AI Sandbox briefly last week - the latest iteration of Meta Advantage. It’s essentially a testing playground that runs ads past thousands of potential customers across Meta’s social media suite to determine the most effective campaign.
By analyzing CTRs, view time, and passive conversational data, the AI Sandbox uses generative tech to craft the perfect advertisement.
Using pure data, the AI can determine which graphic would capture the most attention, the most effective social platform to promote it through, and the best audience to market the advert to.
THE LAST BYTE
SEO is dead.
It’s one of the phrases guaranteed to elicit a reaction from most marketers. And after many predicted deaths, this time AI is the shadowy figure with the scythe.
The staunch “SEO is immortal” believers tell us the number of searches grows year-on-year, with billions of searches happening every day and the $100B SEO industry is increasing every day.
The impact of the AI revolution can’t be denied.
And just this week Microsoft announced all ChatGPT users will be able to access search data while using ChatGPT via a Bing plugin. This means you’ll be able to ask ChatGPT a question and it will include search and web data, and even citations, when it responds.
As the relationship between AI and search becomes intertwined, it is changing how people interact with the digital world. Instead of returning just a list of (SEO-optimized) website links, you now get an AI-generated conversational response as well (e.g. Google) or a conversational response peppered with search data (e.g. ChatGPTxBing).
Now the race is on to discover how to “rank” in AI-generated responses. How much influence will traditional SEO tactics have, if any?
Money talks though. The revenue search companies like Google get from advertising won’t be ignored. While they are keen to keep ahead of the AI curve, they aren’t going to throw away billions of dollars in ad spend.
So SEO-optimization still counts for something.
My bet is It won’t be long before it evolves again though.
See you next week,