Is AI Human After All?

Guy Galloway
July 18, 2024

London is best described as lots of little towns that grew into one larger city, over many years. This is also a good way to think about where we are now with AI tools like ChatGPT.

But first, let’s take a tour of the “little towns” that helped build this larger city.

First, we had hyper-targeted Social Media ads and email-flows. These were based on “what do people want” and used many of the psychological levers highlighted by Cialdini in his book Influence: The Power of Persuasion. But they weren’t human-focused. They were profit-focused: how can we get as many people as possible, who are interested in widgets, to buy my widget?

Then we had “Big Data”, where we took ALL the information we had collected from these targeted ads and CRM platforms and organised it into databases to tell us “if these people like X, then they’ll also probably enjoy Y”.

Again, it wasn’t human-focused. It was business-knowledge-focused: how can we get as many people in these cohorts, who show behaviours that suggest they might like widgets, to buy my widget?

Then we had automated chatbots – simplistic automatons useful for helping inform call-center staff, when they did eventually speak to a customer 1-on-1. Other than that, they were a cost-saving exercise by upper-management and a huge frustration for human customers.

And once again, they weren’t human-focused. They were business-protection focused: how can we get as many customers as possible, to not complain about our widgets and/or buy more widgets?

The Rise of Human-Focused AI

The common theme here is that all these technologies were coming from an out-dated broadcast media mindset of one-way-traffic: we say, you listen (and then buy).

The other issue with all of these technologies is that they were also based on large amounts of averages: whichever thread is the most common is the one we will follow.

We still see this today in PMax campaigns. After a certain number of creative optimisation rounds, the images end up all looking like the same generic image, albeit with a few slight differences because that’s what the machine told us worked best, on average. But humans are all unique, so after a while performance plateaus and we have to start the process all over again.

Which is what makes the current AI models so radical: they’ve learned how to talk like a human, but think like a machine. This means it can give us highly accurate information, tailored to our own individual quirks.

Essentially, AI models, like ChatGPT, are now merging all the above “little towns” into one central, intelligent, *human-focused* capital city. And that’s the key: it’s human-focused. It’s based on the formula: humans + machines > humans alone.

@thesmarthomeguy Here are 3 ways #BigTech is using #AI to make the human experience more real. Watch to the end to hear the most important thing about all this. I just finished reading today’s email newsletter from @The Deep View. And here are are the 3 ways AI is being used to keep it real online: Number 1: X (formerly @Twitter) just launched enhanced security features to better combat AI-generated images and videos on its platform. Basically, if an image of an influential person starts circulating on the platform, it gets flagged as being a fake, generated by AI, all subsequent posts featuring the same image will also be flagged automatically. What’s so good about this? Improving #OnlineTransparency is much needed - we’ve all seen what happens when it gets ignored - making it easier for humans to understand when they’re being misled online is better for all society. What are @Facebook doing to prevent this? Good question! Here’s the answer... Number 2: the AI-tool #FACET has been designed by #Meta to detect biases in AI in imagery and video, particularly when it comes to occupations, activities, and demographic features. Why is this important? Human-based biases get naturally baked into AI and #MachineLearning algorithms. We’ve known this for a long time. This is one measure being taken to reduce this. Good times! And it seems to be working well, as #FACET detected biases in Meta's own commercial #DINOv2 algorithm. Awkward. The next #aiupdate is a game changer… Number 3: #OpenAI is helping educators understand how #ChatGPT can help improve the quality of lessons they give their students. Recently, OpenAI (@OAI) released a guide that highlights the many ways their AI can help improve their educational experiences, opening the doors to richer, deeper academic opportunities. And now finally, here’s the most important thing about all this… The main takeaway for me from all this was how the initial excitement and fears generated by making AI publicly accessible over the last year or so have simmered down and now reality has set in. This is proof that AI is now an everyday part of our daily lives. But we’re still at the very start of this #generativeAI adventure, so it’s really important to keep learning. That’s why I make posts like this, to help you learn about it, too. We’re seeing a huge shift and with that, we need to have boundaries and controls in place to ensure we don’t spin out of control and let our worst fears about AI and machine learning come true. With big tech stepping up like this, it shows they too understand the importance of this and are instead creating ways to ensure AI and Machine Learning is used to IMPROVE our lives. To make it an ally, not an enemy. So what does all this mean? The equation: “humans + machines are greater than humans alone” is more true than ever. So let’s be positive. Let’s learn about the amazing things AI can do and encourage others to do the same by sharing videos like this and comment with what you think on this topic. “Is AI our friend or is it our foe?” I want to hear your thoughts. Follow me, too, for more on AI, #EmergingTech, trends and #smarthome fun. #pythonprogramming #techtok #smarthomeguy #artificialintelligence ♬ original sound - SmartHomeGuy

This is why ChatGPT has exploded in popularity.

It takes people as who they are. It listens to them and adapts to each one’s own communication style. It’s more like a really-helpful, very patient friend, than an obvious robot.

AI’s true intelligence was being able to learn from all these separate intelligences, bring all their knowledge together into one space and understand how to communicate it as a human.

Dive deeper into the AI revolution! Still curious about how the world of AI and human interaction is evolving? Tune in our Paris Talks Marketing podcast episode offering a glimpse into the future of AI assistants and interactive Web3 with Alexander De Ridder!

Guy Galloway

Senior Client Success Manager

I'm responsible for improving my clients' digital marketing performance. I'm driven by new ideas, experiences & teamwork.

https://www.linkedin.com/in/gcmgalloway/