Why I hate ChatGPT content, Mistral releases Le Chat, Google gets into trouble for reinventing history, and is AI crowding out ideas?

Written by Fola Yahaya

Thought of the week: the coming AI content backlash?

I hate ChatGPT-generated content. I can spot the bland, overly formal or perfectly structured rubbish a mile off. This week alone I’ve received three ‘professional’ emails, two sales proposals and a technical report written by an international organisation that were all barely disguised ChatGPT guff. I get it, and especially if you’re a non-native English speaker, ChatGPT is like manna from heaven. Quick and plausible-sounding content for no effort.

However, using (unedited) ChatGPT in your written comms is like giving your audience the middle finger. It shows you ultimately don’t care what they think of you, and that you think your audience is fundamentally stupid. I get the argument that AI writers can be a good antidote to ‘blank page syndrome’. However, many of us are sliding down that slippery slope towards using ChatGPT for all of our writing. This will only get worse as it’s embedded in browsers, phones and even our digital twins.

So stop using ChatGPT indiscriminately. Unless you're protesting against having to write useless content in the first place, remember these rules of thumb as to what ChatGPT is great for and when it should be left well alone:

Just ChatGPT it:

Compliance and bureaucracy

• Form-filling
• Complaint letters
• Parking ticket appeals
• Low-level legalese
• Tick-box exercise reports

Keep it human:

• Homework – Obviously.
• External communications – This includes marketing content. AI-generated content is now so prevalent that anyone who still writes normally will stand out.
• Emails – Controversial, I know, but you can learn how to write effective emails in 30 minutes (keep it short, have a central message, have a next action, only send it to people who need to read it, etc.).
• Personal correspondence – If you can’t be bothered to write something nice to someone you care about, don’t bother!

Is AI making you give up on that business idea?

This interesting article about Hollywood producer Tyler Perry cancelling an $800 million expansion of his Atlanta studio got me thinking about the ‘crowding effect’ of AI. AI is accelerating the ‘death rate’ of start-ups who had a clever idea, only for a tech giant to roll it out as a free feature a few months later. Think copy.ai, Jasper, Blinkist and a host of start-ups that are no longer relevant thanks to ChatGPT doing what they did but better and cheaper. As AI becomes a general-purpose technology, there is a real risk that those who control the most effective generative technologies can pretty much do any task. From automating clicks on a page to generating videos from a single prompt, very few companies will survive on business models for services that can be done by an all-powerful and all-seeing AI.

I have been watching AI very closely and watching the advancements very closely. I was in the middle of, and have been planning for the last four years, about an $800 million expansion at the studio, which would’ve increased the backlot a tremendous size – we were adding 12 more soundstages. All of that is currently and indefinitely on hold because of Sora and what I’m seeing. Tyler Perry

Prompt: A beautiful homemade video showing the people of Lagos, Nigeria in the year 2056. Shot with a mobile phone camera.

Google apologises after controversy over AI image feature

Google’s stock sank 4.5% and everyone and his dog attacked the company because its new ChatGPT competitor, Gemini, created racially inaccurate depictions of historical figures.

Personally, I quite like the images. History has always been a blend of fact and fiction, so why not let chatbots reimagine everything...? But seriously, if within a year it becomes impossible to tell the real from the unreal, how do we confirm truth? The answer lies in, rather boringly, more regulation, authentication mechanisms and holding tech giants to account.

However, this is unlikely to be effective. In the race to stay relevant, Google et al. are in ‘move fast and break things’ mode and are abandoning their role as responsible guardians of our data (and history), so expect more risky releases of far-reaching tech in the months to come.

Stable Diffusion 3 takes the lead in image generation

Prompt used: 35mm realistic photo. A black ethnicity person wearing a reflective vest, holding a clipboard doing a survey, smiling and looking into camera, yellow and teal theme, --ar 16:9 --v 6.0 --style raw.

Stability AI has unveiled Stable Diffusion 3 (SD3), their latest image-generating AI model designed to compete with new offerings from OpenAI and Google. Stability AI, the company behind SD3, is aiming to become an indispensable white-label generative AI platform.

Chrome gets a built-in AI writing tool powered by Gemini, but where are the really useful AI agents?

Google Chrome has introduced a novel AI writing generator, leveraging its Gemini AI model to expand the ‘Help me write’ feature from Gmail to the entire web. That’s all well and good, but when is Google going to bring some real AI to its lame Assistant?

Google Assistant, Alexa and Siri are all fairly useless when compared with the responses you get from ChatGPT and exemplify why people were so underwhelmed by AI back in the day. I cannot wait for Google, or more likely OpenAI, to give me an AI PA (personal assistant). It’s less about making me more efficient and more about doing the tedious stuff that we all waste our very short lives on e.g. filling in PDF forms, being on hold, booking a medical appointment, returning stuff… the list goes on.

DeepMind is coming for gaming!

DeepMind just unleashed Genie, an AI that can create whole interactive video games based on a one-line prompt. Trained on 200,000 hours of video game footage scraped from the internet, all without a hint of supervision.

What’s crazy is that Genie can create games without any instruction. The magic behind this? A cocktail of tech wizardry involving a latent action model, a video tokenizer, and a dynamics model.

But hold your horses, it's not game time for everyone just yet. This groundbreaking tech is still in the lab-coat phase, courtesy of Google DeepMind, so no public play dates (cue the eye-roll memes). And sure, it's currently rolling games at a nostalgic 1FPS, but remember, this is just the starting line. The potential here is bonkers – we're on the brink of a world where anyone can whip up their own digital dreamscapes. Buckle up; we're just getting started in this wild, wild world of AI-generated gaming.

In other news…

• Klarna, the Swedish buy-it-now, pay-later giant, says its AI assistant does the work of 700 people. This comes after it laid off… 700 people in 2022–2023, citing high inflation and the war in Ukraine as the reasons.
• Mistral, Europe’s only competitor to OpenAI, partners with Microsoft and releases the Gallic Le Chat.
• The research is in: when we outsource the dull but important stuff to AI, we forget how to do it ourselves.
• Audio-to-video just got really good. Alibaba just dropped a scary vision of the future. Combine this with Sora-like rendering of videos that can be fully lip-synced with any audio, and we can be taught by Einstein, tutored by John Coltrane and historical exhibits can be totally interactive. On the downside, it continues to threaten actors and truth itself.

AI tools we're playing with this week

Suno.ai - An AI music generator that just gets better and better. Surely a pop-music killer? If Spotify buys this, the majority of new songs will likely be algorithmically generated. The below snippet took literally 30 seconds to make and is based on the prompt: “a pop song about AI”.

ZipWP - Lets you create an AI-powered WordPress website in seconds.

That’s all for this week folks. Next week we’ll be bringing you more AI news, ideas and useful tools to support you during the AI transition. Subscribe for the latest innovations and developments with AI.