What does GPT-4o do? What does it mean for society?

Have you heard about Open AI’s GPT-4o yet? It’s the latest and greatest development to the beloved ChatGPT. And it’s a pretty big update indeed, for the tech space and society alike.

It’s the dawn of a new era once again for OpenAI (and the rest of us), with it’s new release of GPT-4o.  The ‘o’ stands for Omni which isn’t scary at all, right? This latest update is available for anyone and everyone to use via ChatGPT. So let’s dig into the details, plus a little bit of opinion-spice from your friendly, neighbourhood IT company.

What does the update include in GPT-4o?

So basically, GPT-4o can “reason across audio, vision, and text in real time” according to the OpenAI website. If you just watch the videos of people interacting with it in the demos, you’ll understand what OpenAI mean. In a nutshell, it means that you can practically have a video call with it. Except you’re the only one on camera, and GPT-4o is a series of little, moving black dots. Although, we’re sure it won’t be long before it gets its own visual identity.

When you see it for yourself, you’ll probably be thinking ‘crikey, it’s like interacting with a real human being’. Cue everyone’s brains going wild with the possibilities that this kind of technology could pave the way for.

This is the video of OpenAI introducing GPT-4o…

Here’s what OpenAI are saying about the update…

“GPT-4o (“o” for “omni”) is a step towards much more natural human-computer interaction—it accepts as input any combination of text, audio, image, and video and generates any combination of text, audio, and image outputs. It can respond to audio inputs in as little as 232 milliseconds, with an average of 320 milliseconds, which is similar to human response time in a conversation.”

So you can have a conversation with it. And not just over text (anyone remember messing with SmarterChild in the old MSN days?). You can one-way video call with it. It can respond to your appearance, your facial expression, the background, your tone of voice. If you speak to it, it’ll speak back. And the thing is, it has a surprisingly and eerily likeable personality (if you can call it that)!

Have OpenAI taken the human likeness too far?

Wellllll. That’s a matter of opinion. We’re sure OpenAI don’t feel that way. But the more you find out about GPT-4o, the more the human likeness will surprise you – that’s for sure. For example, GPT-4o can use sarcasm. And even interact and sing with other GPT-4os! Seriously, you gotta watch the videos.

But the first thing that’ll strike you when you witness it for yourself is the uncanny resemblance to the intonation of a human voice. The rise and fall in pitch when GPT-4o speaks to you sounds just like a real human. And a really nice, bubbly one at that! It’s a new realistic sense of human-ness that people are just not used to. So many of our reservations as a society probably stem from this inherent sense of ‘other’. The creepiness that stems from that feeling of being in ‘uncanny valley’ (as explained by the online icon, VSauce, 10 years ago, in his video ‘Why Are Things Creepy?’).

Although we’ve not had the opportunity to interact with GPT-4o ourselves in this new ‘AV’ way, we imagine it might feel the same as when you apologise to a table after accidentally bumping into it. Or when you thank the machine that prints your parking ticket out.

Like… Are you meant to say please and thank you to GPT-4o? And if you do, or don’t, does it do anything? Does it mean anything?

Although speaking to GPT-4o could be seen as a small and meaningless interaction, these are the questions that spring to mind and they relate more to the human condition and our sense of Self than the technology itself. So really, it’s up to you whether it’s all gone too far and it’s time to go and live in the forest.

How will GPT-4o change things? *eerie music*

Well, there are some very obvious good things, and very obvious bad things, that this type of technology can be used for. Let’s start with the juicy bad stuff.

What happens when an AI can replicate someone’s voice (which it can already) – but also interact with that person’s loved ones in a totally organic way with no delay? Social engineering cyber attacks and scams go through the roof, and innocent people’s emotions, vulnerabilities and good will are the victims. That’s what! Or how about accessing someone’s accounts over the phone; ever had to say “my voice is my password” to the bank’s automated response voice on a call?

On the other, much less doom n’ gloom side of things, is the good stuff. Customer services could improve dramatically. Everyone hates those automated responders, right? But what if you could explain yourself over the phone without the fear of sounding like a robot yourself? And for the real robot on the end of the line to understand you properly!? It’d be a miracle. It could also mean that call queues don’t exist anymore, as GPT-4o could deal with an infinite number of people, at any time of day or night.

And now it’s time to address what our weird, little minds think the elephant in the room is: What if someone falls in love with GPT-4o? This kind of technology could be a lifeline for people who are lonely, and we all know that social isolation is practically an epidemic. But how far is too far when someone wants to marry it? It’s just food for thought. We’re not sure anyone really has the answers for this stuff.

How to access GPT-4o?

Everyone wants to have a go. There’s no shame in it! Well, for now, you can access the text version of GPT-4o through ChatGPT as usual. All the snazzy stuff we’ve been talking about in this blog post is gradually rolling out over the next few weeks. So you’ll have to sit tight and wait for the cool stuff!

But in the meantime, if you haven’t used ChatGPT at all yet, there’s genuinely a whole host of interesting things you can do with it. And loads of it is helpful, especially at work. It can create marketing content like social media posts; give you ideas on pretty much any concept you can think of; reconsolidate data and information; create travel itineraries, meal plans, how-to guides… You name it! Click here to give it a go (it’s free).

If ChatGPT’s not your thing, but you’re interested in exploring the world of AI, check out our other blog post Free AI tools for boosting productivity at work. It covers 5 free, online tools that could be handy for you. Work smart, not hard, eh?

Got a question? We can answer it. Click here to get in touch.