![](https://strategicagenda.com/wp-content/uploads/2024/09/Strategic-AI-Template-banner-768x432.png)
Written by Fola Yahaya
The last 12 months have been a roller coaster ride for writers. On the one hand, AI has clearly made writing faster and idea generation almost instant. However, it has also seen a massive increase in AI-generated e-books that are starting to seriously muddy the waters. Countless sites (and YouTube tutorials) enable anyone and his dog to automate the ‘tedious’ process of prompting an AI chatbot to write a 200-page e-book, complete with cover art and back blurb, in under a minute.
This has led to a deluge of AI books that rip off existing works and/or just contain bullsh*t content. It has got so bad that Amazon has decided to limit the number of e-books that authors can submit.
Why AI-generated content is a problem
Despite the range of forces that threaten the publishing, sales are steady (around 1 billion a year in the US alone) and physical books are maintaining their post-pandemic resurgence. So whilst AI might enable more bad books to sneak into bookstores and pollute Amazon’s Kindle store, discerning readers will continue to vote with their wallets and quality will out.
Anyone who knows me knows that I am an Apple sceptic. I don’t like their constant upgrade cycle and the lemming-like behaviour of Apple fanboys and girls. However, this week Apple blew everyone away at their annual Worldwide Developers Conference (WWDC) in Cupertino. It was at this event that Steve Jobs famously launched the iPhone, a product that revolutionised smartphone usage and that also revived a company that was on its knees. So, expectations were high and boy did Apple deliver.
Dressed head to toe in Silicon Valley black, Tim Cook strode onstage to wild applause. Silently, he then sat and waited whilst two Apple workers placed what looked like a grey circular disc emblazoned with the Apple logo centre stage. Cook then stood up and moved back on to the spotlighted disc and, like a big tech version of David Blaine, floated slowly upwards until he was floating 10 metres above the disc.
And there he remained, motionless, for exactly 60 seconds, as the audience gasped, cried and some even blasphemed, before he finally broke the silence and uttered what will surely go down in history as the punchline to the greatest product launch in history:
“Ladies and Gentlemen… I give you the iHover.”
Ok, I made up all the above. It took me all of five secs and a single line prompt to create this on Midjourney. Why? To show how easy it would be for me to spread a false story on X or TikTok about… well, anything.
A big Apple WWDC event is taking place this week but, despite the hype, it’s a bit of damp squid. The TLDR (too long; didn’t read) take on the event is that, after giving up on building its own AI, and flirting with every major provider, including arch-rival Google, Apple have finally settled on OpenAI as the AI engine that will be embedded locally on your next iPhone.
Siri will now basically be powered by a limited version of ChatGPT, and there were a bunch of updates to Apple’s $3k VisionPro VR headset and some other underwhelming stuff that you can find out here if you have time to waste.
My take on it is that Apple is struggling to stay relevant, having abandoned its Apple Car project and given up on besting OpenAI and Google in the AI race. But this doesn’t matter, as just mentioning AI bumped up its share price by 7%!
Everything you’ve ever posted online has (probably) been used to train AI
Though Microsoft has decided, at least for now, to shelve its recall feature – you know, the one where your PC silently spies on you and learns from your actions – you can bet that every single company that has your data is already mining it to train its AI for eventual profit. This week it was the turn of Adobe to change its terms and conditions so that the software giant can use all your lovely creations, even unpublished and in-progress projects, to train its AI models.
The updated terms that have been in effect since 17 February 2024 mention that Adobe’s automated systems may analyse user content and fonts using machine-learning techniques to improve services and software.
Meta has also changed its T&Cs to similar effect, so all your Instagram and Facebook posts from as far back as 2007 will now be used to to train artificial intelligence tools.
So if you were under the misguided illusion that big tech cares about your precious data rights, think again. Everything you have ever posted online is or soon will be mined to train AI systems to sell you more stuff.
Killer tips to up your ChatGPT game
It’s very easy to get stuck in an ‘all AI is bad’ gear without celebrating what it’s so great at – writing inconsequential content. If, however, you find your AI bot is all too eager to spout out a stream of consciousness reply, then follow these handy tips:
- Use the phrase “Be precise”. Tack these two words on to the end of your prompt and sit back and watch as the bot delivers a something short and sweet rather than long and languid.
- Upload examples. LLMs are stupid, so upload a sample document so that they know exactly what the output should look like.
- Get it to browse the internet. Now that the free version of ChatGPT and AI search engines like Perplexity can access the internet, you can paste URLs and get the bot to grab examples or source material from specific sites.
Why your next teammate may be an AI bot
AI agents have been the hot new thing for most of 2024 but it’s only now that we’re starting to see them applied to business workflows.
Asana, a project management tool, is trying to stay relevant in the AI era by introducing a bot that can act as a teammate. Clearly the direction of travel is not just AI teammates, but AI supervisors. For example, Uber’s job allocation AI already plays this role by deciding which drivers will attend to your request and when to exploit people with surge pricing.
Asana bills their new AI teammates as hyper-efficient, ever-present collaborator overlords – a feature likely to become standard on all platforms as the tech becomes more widespread.
Kling AI is the new Chinese-developed AI text-to-video generation model that is filling the vacuum left by not much Sora news. Kling AI leverages advanced 3D reconstruction technology to produce vivid, lifelike videos that are up to two minutes long.
It will be some time before it makes financial sense to use AI for videos. There are so many moving parts (actors, music, lighting, scripting to name but a few) in a video that there would always be substantial editing, and text-to-video generators can’t (yet) produce a consistent output. But it will continue to exert downward pressure on prices as buyers mistakenly think video can be created quick and cheaply.
That’s all for this week. Subscribe for the latest innovations and developments with AI.
So you don’t miss a thing, follow our Instagram and X pages for more creative content and insights into our work and what we do.
Network Hub, 300 Kensal Road, London, W10 5BE, UK
© Strategic Agenda 2024. All rights reserved. Privacy policy | Sitemap.
We’ll send you quarterly updates so you can see what we’re working on, what the future holds and how we’re shaping it.