Before this was AI, we called it something else

My brother in christ, that is literally just a featured snippet

I don’t think technology is as clever as we’re being told it is.

Something happened to me last year that borderline radicalised me on this subject. We were on our way home from our holibobs in Fuerteventura. The airport has passport control after security, before you get to your gate. Passport e-gates. You know the ones. Pop your passport in, look yearningly into the camera, and some sort of advanced technology scans your face and the passport photos. Then, a weird algorithm beyond our comprehension figures out if you’re both the same worldly being by measuring like, pupil distance or some shit.

At least, I thought that was how they worked.

On the other side of the gates, there was a man in a lil booth. He was there because us Brits now need to have our passports stamped whenever we go to Europe. I gave him my passport with my best ‘buenas dias’ and he flicked through to find the right page. While I was waiting, I noseyed at his screen for a sec. And what I saw shocked me to my core.

Or, to be slightly less of a diva about it, I was surprised. This dude had a live feed of all the e-gate cameras. A security measure, maybs? But when people scanned their passport, it popped up next to the feed from the cameras. He had a bit of a think, then pressed a button. The e-gate opened and people joined the queue to have their passport stamps.

These machines were not powered, as I had assumed, by some infinitely clever biometric technology. They were powered by this man in the booth. The e-gates simply saved him a little bit of time, so that instead of scrutinising our faces one by one, he could see six people at once.

It made me ask a lot of questions. And I am an individual who, at the best of times, is asking at least three to five questions at any given moment.

Is this how all e-gates work? Is that why you’re sometimes randomly waiting ages before it works? In every airport in the world, is there a little gremlin in a hidden room somewhere, looking at a sea of tired faces and scans of passport pages and passing judgement?

And why did I not know this?

My current theory, beyond this being one weird anomoly on a distant Canary Island, is we’re being led to believe technology is amazing and incomprehensible to the average mind. When a lot of the time, it’s just something we’ve always had, but with a new hat.

Which brings me nicely onto AI.

I don’t think AI is as clever as we’re being told

There’s a lot riding on AI, which obv creates a sense of palpable desparation amongst its extollers to make it work. At all costs.

That’s led to quite a lot of things “becoming AI.”

Things that, a few years ago, were simply… just things.

I’ll explain with a few examples.

Featured snippets are now ‘AI summaries’

Google et al are waxing lyrical about their AI summaries. Search for a question, AI finds the answer and pops it right at the top of the page, so you don’t have to navigate any further.

Except, Google’s been doing that for years. And they called it “Featured Snippets.” Below is a screenshot from 2018, but the article I took it from explains the feature was around in one guise or another for years before that.

I’m old enough to remember when every SEO team under the sun was hyperventilating about this feature taking away their web traffic. Just like they’re now doing the same with AI summaries.

Same feature, new hat.

Now, I wanna caveat here, I actually asked Google what the difference is between AI summaries and Featured Snippets. And it was very QUICK and very CLEAR in its answer that they could not be more DIFFERENT. The snippet is text taken verbatim from one webpage and presented as an answer. AI summaries read lots of webpages and write a new answer based on the findings.

In a way, fair. But I can’t shake this feeling that the underlying technology, the thing that finds the right information and regurgitates it back up for you like a loving mother bird, hasn’t changed as much as we'd think.

Grammarly

When it comes to proving that AI is just some crusty old algorithm tarted up by an instant makeup gun, you really can’t get much of a better example than Grammarly.

It now calls itself “Free AI writing assistance.”

Because it can take what you’ve written and ‘use AI’ to re-write it, change the tone and so on.

But I distinctly remember being in the office, pre-COVID, and seeing everyone and their mums on Grammarly, pasting in things they’d written and asking it for suggestions. It was basically… the whole idea of the tool.

God bless Wikipedia, which says: “In 2019, Grammarly added a tone detector to its writing assistant. This tool uses set rules and machine learning to help users gauge the character of their writing and tailor it to a particular audience.”

So, at least three full years before ChatGPT blazes on the scene to change writing forever (lol), Grammarly is using a sophisticated machine? algorithm? thing? to detect the tone of what people have written and give suggestions. Which is basically 99% of people’s reason to use AI like Chatty Geepts in the first place.

Obv, Grammarly is now very proud of how what it does is ‘powered by AI.’ Especially as it’s just been sold and rebranded to Superhuman, presumably for gazillions of dollars.

But, at its core, is it not still just doing what it’s done, for years and years? Just in a slightly more sparkly interface…?

Voice assistants

Science News Today gives its #1 use of ‘AI in Everyday Life’ to be smartphone voice assistants.

“When a person asks their phone to set a reminder, check the weather, or send a message, the AI system converts speech into text, analyzes the meaning of the sentence, determines the user’s intent, and executes the requested task.”

So, pray tell, what has our Siri been doing in the years between 2012 and, say, 2022, when AI started going into everything? Was she an early example of AI?

Probably not, because if she was, wouldn’t Apple have made much more of a song and dance about it? “We discovered AI a decade ago, suck it Altman,” sort of vibes??

Siri does what she has always done. But now, because she speaks and comes up with answers, she is AI.

What did she do to deserve this, TBH.

Loadsa things

To save us all some time, let’s just say I can go through the entire aforementioned Science News Today article and grumpily shout “But that’s ALWAYS been a thing!?”

Personalised recommendations on streaming platforms. Yep, we definitely had that before the pandy. Google Maps (really?? We are saying Google Maps is now just an AI thing??). Email inbox filtering. Facial recognition. Online shopping recommendations. Banking fraud detection. Language translation. Smart home devices. Predictive text and autocorrect, for crying out loud. Chatbots. I remember being annoyed about having to discuss them at length in about 2018 or so, in my social media agency days. And even if they are now powered by AI, they still appear to be just as useless. So what really changed?

What now?

Has anyone else noticed this? This continual, persistent re-badging of mundane tech features, that we've had for years, as ‘powered by AI?’

I’d love to know where you’ve spotted it in a software or platform you use a lot.

And I suppose that leaves us wondering, what do we do about this?

In reality, there’s not much we can do. Apart from remember. Remember what we had before, and how it’s changed (or not, in many cases). Remind people that their clever AI agent is just a load of old algorithms sellotaped together.

(And if that’s the case, what’s the deal with these humongous data centres that are sucking up people’s tap water like that scene at the end of There Will Be Blood?)

Now, a final caveat. It’s totally possible that I’ve simply grievously overestimated what AI is, what it can do, and how it works.

But I’ve got a really funny feeling that that’s by design. By keeping the whole industry obtuse and mysterious, AI companies keep us in a continual sense of wonder and admiration at this god-like tech they have created. And, when the time comes, they’re probably going to ask us to hand over a hell of a lot of money to use it.

And at that point, we have to remember that we could be paying above the odds, both with money and resources, for a nice Spanish man in a booth, pressing buttons and being called the future.