The week Google's AI answers became Google's problem

Episode transcript

Amelia (00:16)
Hi there, I'm Amelia.

Paul Wood (00:18)
I'm Paul.

Mayna (00:18)
And I'm Mayna.

Amelia (00:19)
And this is Fin the Week. How's everyone doing?

Paul Wood (00:23)
Pretty good.

Mayna (00:24)
Good, thank you.

Amelia (00:25)
Nice to have you back, Mayna, it's been a while.

Mayna (00:27)
Yeah, it's great to be back.

Amelia (00:29)
And we've got a really nice show today. So we're going to be taking a look at some recent industry news. We might even make this a regular feature if it goes down well. We've got five recent news stories that we think are worth having a look at in a bit more detail. And yeah, we'll just have a little bit of a chat, see what we think. So should we dive into the first story? Paul, what is it?

Paul Wood (00:54)
Yeah, so the first one is something that is potentially quite big news, but might have escaped some people because it wasn't really, as far as I can see, widely covered. Essentially a German court has reviewed Google's use of AI overviews. So for those who aren't necessarily too up to speed on that, AI overviews are when you use Google search and you get that chat-style interface that appears at the top of the search result that gives you a rundown, basically gives you the full answer to what you've asked. So that is an AI overview. But a German court has reviewed that and has determined that Google's AI overviews are essentially Google's own words, and therefore Google is liable for anything that it says that's wrong.

So the article that covers this, which is from the Decoder website, has said that AI overviews generate independent, new and substantive statements in Google's own words. So Google owns them. And one of the key things that stood out for me was it said only Google can check these statements in advance of them being published, if you like. And therefore Google is liable for them. And unsurprisingly, Google took a different view. They argued that users know that AI answers aren't necessarily accurate, and they know that they can click to check the source of a piece of information to verify it themselves. But the court argued otherwise. And actually the article that I've been referencing itself referenced a piece of research from the Pew Research Center that said only 1% of users click to the source of a piece of information in Google's AI overviews. And actually this relates to another story we're going to go through in a minute. So yeah, basically very few users actually click through. They just take the answer Google gives them and kind of leave it there.

So yeah, I think this is quite big news because it sort of sets a precedent, and it's something we've covered in previous episodes — the concept that Google has turned from being a kind of modern day yellow pages that just points you to something into what it is now, which is a sort of publication in its own right. You ask it something, it gives you a fully formed answer. And this court finding is saying that that means Google is liable. And that's particularly relevant for an industry like ours, financial services, where information that is published often has to go through a compliance check. And it's going to be interesting to see how Google handles this. But yeah, I don't know what the view of you guys is, because I certainly had thought this was coming. And it's interesting to see a court has sort of found this now.

Mayna (04:11)
Yeah, I find this really interesting actually. And Google's sort of response to the situation by saying, well, users can check for themselves, they can look at the sources — and then research showing that only 1% of users do check sources — I think that kind of defeats the whole value proposition of an AI overview. It's like, you don't have to click through, that's sort of the purpose of it. And then that contradicts the idea of Google saying you should always leave to verify sources. So it's definitely an interesting one.

Amelia (04:42)
Definitely. I think, you know, we've discussed this before, it comes back to that element of responsibility, doesn't it? Because it's all very well us having this conversation, but I think of, you know, one of my older aunties going on to Google and Googling something. I don't think it would even occur to them to check a source. They're Googling it. That's what Google is.

Amelia (05:06)
You've already got that kind of, it's coming from Google, I think this is valid — why would those people check those sources? And Google know this, don't they?

Paul Wood (05:14)
Yeah, exactly. It's quite cynical for them to pretend otherwise. I think they know. Because Google is a massive brand, and somebody who isn't in the industry that we work in isn't even going to question the authority of that brand. They're just going to think, it's on Google, I can trust it.

Amelia (05:23)
It's a bit gaslighty, isn't it?

Paul Wood (05:43)
So for Google to say it's not us, we're just compiling stuff from other people — I think this court ruling kind of says that that argument doesn't really wash. And I suspect it will spread further from here, it will become quite a common topic. And it also, it's not just Google that this will impact, I think, because the same logic applies to all AI chat interfaces. ChatGPT, Claude, any of these sources that are presenting an answer will either have to find a way to make it explicitly clear that they're quoting from something and it's your responsibility to verify that, or, as you say, Mayna, they lean into that value proposition that they present, which is that you come here and you get all your answers — but then they're going to have to take ownership of what they say.

Mayna (06:35)
Yeah, no, I think that's a good point. And there was some research done that shows that Google's AI overviews, their answers are only correct 91% of the time, which sounds great. But if you zoom out, that's still millions of wrong answers every hour. And if you zoom out even further and look at all the other AI platforms, that's still a lot of misinformation or hallucination.

Mayna (06:59)
Yeah, the scale is quite scary. But I think AI's still so new, isn't it? So the larger companies are still working their way to find the perfect platform and solution.

Paul Wood (07:12)
Yeah. And I wonder if this could be the start of a new... So, years ago in the world of SEO, there was a concept of YMYL, so your money, your life, where Google was kind of explicitly quite careful with any information on very important protected topics. So money and life kind of topics, essentially money and health. And so the way the algorithm worked and surfaced search results was slightly different. It sort of had a higher bar for results to get over. And that will still apply today with general search results. So these AI overviews will be drawing information from search results that bake this kind of process in.

But I wonder if this kind of court ruling ups the ante with that and says, well, you're probably going to have to go a step further now, and topics that are protected, you're going to have to be really careful with what you're essentially publishing. An AI overview can't just generate something based on its training set, because there might be a flawed piece of information in there, or it might jumble up two different sources of information and create the wrong answer. And when it comes to money and health, they're two areas where you can't mess around. Firstly, because you're going to cause problems for the end user. Secondly, because there are a lot of well-paid legal experts keeping an eye on this. So yeah, I think this is definitely one to watch, and I suspect we might see more of this type of thing over the coming weeks and months.

Mayna (09:15)
It definitely raises a protection gap. Because say if a company — like the two cases given in the German court — are affected by misinformation from AI, that company can't sue source websites, because that's not the words of those websites. But then Google will say, well, actually, it's not our words, it's words from third party sources. So it's just this back and forth, isn't it? It's still a bit unclear where the liability lies.

Amelia (09:46)
It'll be interesting to see where that lands. Moving on to the next story. We're sticking with Google, aren't we, Paul?

Paul Wood (09:51)
Yes, yeah, and this is somewhat related, but it concerns a study carried out by a company called SparkToro. Anybody who has worked in the search industry in the last 15 years probably knows of a person called Rand Fishkin, who is a founder from Moz. But anyone who hasn't worked in the search industry probably won't know him and won't know the company. But he's seen as a bit of a godfather of search research, and so his new company SparkToro looks at sort of audience research, and he has been big on the topic of zero-click marketing, as he calls it. So this concept that fewer and fewer online searches result in a click, and he frequently publishes research in this area.

So the kind of piece of news we're looking at now is the most recent piece of research into that. The finding is that fewer than one in three Google searches now leads to a click of any sort. And potentially more interesting than that headline stat is that, of the times that somebody does click on a search result when they search, only a portion of those — let me just check — yeah, only 66.61% go to what you would call the open web. So that means that a kind of small portion of all searches goes to the open web. The rest either don't click, or they go to another Google platform. So it could be Google Maps or YouTube or a Google business profile, something like that.

So the sort of key takeaway is that if you're seeing a decline in traffic, particularly from search — and let's be honest, most websites get most of their traffic from search and particularly from Google — so if you're seeing a decline in that, it's not just you, this is an actual trend. So the numbers have gone up since they last reviewed this. 68% of Google searches ended without a click. And as it happens this morning — so we're recording this on Wednesday, this episode will go out on Friday — this morning a new update has been published on the SparkToro website that breaks it down by different countries. And the UK is currently the worst in terms of zero-click searches. So only 30% of searches result in a click. So yeah, this is a huge thing. And of those 30% that result in a click, 76% go to the open web. So I can't figure out the maths from that, but whatever 76% of 30% is. Here we go. For every 1,000 UK Google searches, 232 clicks go to the web. So yeah, that's what you're working with.

Amelia (12:54)
Are you surprised by those numbers?

Paul Wood (13:08)
No, not really, but that's because I've been following this story for a long time. But I think people will be surprised, because most people who aren't obsessed with this whole world will just be used to using Google and searching, and in their heads they still search and click. But when you stop and think about it — well, actually, throwing a question to you, Amelia. Do you still feel like you click on search results, or do you find yourself increasingly searching and then going nowhere else from there?

Amelia (13:44)
Yeah, I find that a lot, actually. I do live radio, so I find a lot of the time when I'm on air, it's so quick when I want to check things. When I just want to check something, you know, I obviously do a lot of fact checking for music, for example. And that's something that before would be a bit of a process, and now it's a five-second thing — I don't have to go beyond the first page of Google. But even so, these numbers seem crazy to me. What do you think, Mayna?

Mayna (14:14)
I think this trend of declining clicks — AI overviews have been sort of the engine behind that acceleration. I feel like a lot of companies who aren't immersed in this trend like we are, they sort of see it and start to panic. It goes back to the concept of, this doesn't mean the end of search, it just means search is changing. And businesses have to change to adapt to the new way people are finding information. Which I think is also only the beginning. Like, AI overviews are still pretty new and they don't show up in all searches. I can't remember the exact statistic, but they're still not rolled out completely. So yeah, this is only the beginning. I think the future's still going to be very interesting.

Paul Wood (14:44)
Yeah. And as well as that, I think we're going to see a migration away from people using classic search engines and towards AI chat applications and things. Or, alternatively, the classic search engines are increasingly going to look like chat interfaces, which is another sort of theme we've discussed multiple times on this podcast. But I think, whilst the stats are pretty sobering, I suppose, it represents a big shift. I don't think it's all bad. I think it just changes things. I think it changes the role of a website, because a website used to be the place you expect people to end up every time, regardless of which stage of the user journey they're at.

Whereas now a website's role is sort of multifaceted. It partly exists as your sort of source of truth that trains an AI chatbot or an LLM, provides all of the information an LLM might need to then compile answers that include you. And then later on down the line, it acts as the sort of place that your customers or nearly-customers come to to verify information and to complete a process. But what it does mean is that you're probably not going to see huge volumes of traffic generated from search. It's probably that you're going to use your website as a place to publish your position on a subject and hope that AI references you.

And then the research that SparkToro have done sort of offers up some thoughts. And one of the key things it says is that you need to replace traffic as a KPI in your marketing efforts and start to build what they've called a correlation dashboard, which is quite a smart way to look at it, I think. Because it's basically saying, let go of all of the specific metrics that you used to monitor — and again, we've discussed this before, clicks are disappearing, visits are disappearing — but what you can track is sort of across the board: what are we doing as an organisation that counts as marketing? And then at the other end, what's going well? Is our revenue climbing, or our sales climbing? What's our customer feedback like? All of those sort of correlation metrics. And you shouldn't really work towards the kind of this-in-equals-this-out. It's more like, this is the activity we're doing, this is how well we're doing, and you're just trying to optimise the two together. Which is probably going to be quite uncomfortable for some organisations that have got really hooked on specific metrics. But yeah, that's how I kind of see it. I don't know, Mayna, you do a lot of reporting for us, and I think you've probably seen the impact of this over the last couple of years of how we're sort of tweaking what we're looking at.

Mayna (18:21)
Yeah, measurement and reporting has just changed completely. And we're looking at sort of updating our monthly reports for clients to better reflect how search is working and how users are finding websites, because you can't attribute sources and traffic as easily anymore. Yeah, so, changing.

Amelia (18:42)
It is all changing. I mean, that seems to be the theme of everything we're talking about — change, and I guess rolling with the punches. And it's those businesses and organisations who are open to it who presumably are going to do the best long term.

Paul Wood (18:55)
Yeah. And I personally think it's quite exciting, because I've always felt restricted somewhat by — when you're doing search marketing, you often have to justify every decision you make by, well, what traffic is this going to deliver? And what end result is that going to deliver? And of course that's important, but the conversation never turns to, well, how's this going to impact our brand and our revenue in the long run? When, because that's actually the important stuff. Clicks and traffic is just a means to an end.

And I think it'll drive more creative thinking and more creative campaigns, where you won't have to spend days on end justifying, well, here's where the click's going to come from. And instead you can kind of come up with a really good campaign idea and say, this is the impact it's going to have on our brand, and we want to get eyeballs on this, and this is going to last us for a long time. It makes the whole concept more exciting, I think, and moves it closer to the sort of classic world of out-of-home advertising and stuff like that, where the creative and brave and bold campaigns tend to win. So yeah, I think that's quite a good thing for the industry.

Mayna (20:20)
It's definitely exciting. Like, Paul and I have been looking at automation work recently. And I've been in the industry for four and a half years, and I feel like that was a good amount of time to sort of learn the basics of DM and build on those skills. But now, for the first time in a few years, I feel like I'm learning something completely new, like from scratch, which I think is really cool.

Amelia (20:47)
It seems like a good place to leave that one. It's good — good positive reactions to this. But what will be the reactions to our next story, Paul? So this is about the ChatGPT advertising pilot.

Paul Wood (20:52)
Yeah, so this one comes from a financial promoter website — so an organisation that publishes updates on the kind of financial marketing world. And it's about Hiscox Insurance, and they've joined, as you say, the ChatGPT advertising pilot via Dentsu. So one of their agencies is one of OpenAI's test partners. But essentially, yeah, they're going to be testing running ads on ChatGPT. This one's interesting, I think, for a couple of reasons. Firstly, because it's going to be interesting to see how ads can work in an AI chat interface, because the appeal of an AI chatbot to a user is that it, in theory, gives you a very personalised answer. So how do you integrate advertising into that without it being obnoxious, and keep it helpful?

And I think, for all of the stuff we maybe level at Google, you would have to say that for the most part, Google Ads has been a success, because it generally is on topic. It's generally not too invasive. People know it's there. It's sometimes helpful, it's sometimes not. It's fairly easy to ignore. And it'll be interesting to see how ChatGPT sort of tries to tackle this one. But it's also interesting because — we've discussed this previously on the podcast as well — are ChatGPT, or is OpenAI, the founders of ChatGPT, jumping the gun a little bit? Because whilst they're currently the most used AI chat out there, they're certainly not the nailed-on market leader in the same way that, well, Google is in search. And there's already data showing that Google is clawing back market share with Gemini. And things could change quickly — this time next year we could be looking at a world where Gemini is really catching up.

And then you have to say, well, is running ads really the right thing to do right now for ChatGPT, or is it just going to kill off more users? Because the other bit of news — isn't covered in this, but is relevant — is that Google have said they're not going to run ads in Gemini for the foreseeable future. And I don't believe for a second that that's going to be permanent. They'll definitely change their mind at some point. But that's obviously them saying, well, it's too early for us to say we've cornered the market now, we're going to squeeze the money out of it. So yeah, I don't know, this could go either way for ChatGPT. But the news story is interesting anyway, because the article quotes Stuart Mahoney, who is head of acquisition marketing at Hiscox UK. He has said that the pilot will help them understand how an insurer can exist in this AI-driven discovery world. It will see how this sort of advertising can influence customer journeys and insurance purchasing behaviour.

But the question it raises for me is, who's responsible for ChatGPT's sort of organic information, if you like? So the non-advertising information that appears alongside your adverts. And it comes back to that first story we were talking about, which is that if ChatGPT is synthesising answers that aren't really going through any form of compliance, do you want your ads showing up alongside that if you're an insurer? So that's the biggest risk, I think, for me. But I'm interested to watch how this one pans out, because I think it will be useful to know — it will probably tell us the direction of travel that the industry is going in.

Amelia (25:23)
What do you think of this? Any thoughts?

Mayna (25:25)
Yeah, I kind of view this with a bit of an open mind. I mean, this is one of the first financial services brands publicly putting its name inside a chatbot. And for financial services, such a heavily regulated industry, seeing an insurer step into such an unproven, barely defined ad environment — I think someone has to do it. And, like it says in the article, search marketing is such a critical part of user acquisition, it is important to understand search behaviour. And I think that's what they're doing with this. So my view is, someone has to do it.

Amelia (25:49)
So it'll be interesting to see, as you say, what happens. And with all of this, it's also such early days. What's our next story, Paul? This is about AI citation ranking factors.

Paul Wood (26:15)
Yeah, so this is kind of big business at the moment, because everybody wants to know, how do I rank in AI search results? So this is another person whose background is from the company Moz that we mentioned before. So it's a person called Cyrus Shepard, who now sort of works as, I think, a consultant. But anyway, what he did was he looked at 54 different studies and experiments and explainers on the topic of AI citations, trying to figure out, across all of these different studies, what do they find actually works? What makes one brand or one piece of information more successful at showing up in AI search results than another?

And so his findings were quite interesting, because first of all, it found that if you're winning at SEO, so classic search marketing, then chances are you'll be winning in AI search as well. Which kind of feeds into this whole story that Google is pushing, that, you know, AIO, AEO, GEO, whatever you want to call it, is fundamentally the same as SEO. And I think that is true for the most part, but there were definitely some caveats that this article highlights as well. It's worth mentioning that the article is sort of all about correlation. It's not a confirmed rank — the stuff it lists out aren't confirmed ranking factors. They're just things that the top results tend to share, sort of qualities that they tend to share.

But yeah, stuff that stood out: URL accessibility. So making sure that the content on a website is accessible to a bot or a crawler, as well as a human user. This is particularly relevant at the moment actually, because I stumbled across an issue with a few websites last week that I discussed with a colleague to do with Cloudflare. So Cloudflare is a service that a lot of servers will have installed on them, and it exists as a way to protect against sort of attacks to the website, stuff like that. So it blocks certain bots and crawlers to stop, like, a DDoS attack, where somebody sends tons of requests to a server to basically grind it to a halt. So Cloudflare is used for that sort of thing. But what I found was that Cloudflare was also blocking certain AI bots, which you actually want to give access to.

So there's a bit of work there to optimise that on sites, and it's worth it if you're running Cloudflare or anything similar — do a bit of testing, because there are a lot of new crawlers and bots out there that you wouldn't necessarily have left your server open to before. Because in the past you'd have had Googlebot and different versions of Googlebot, which come to your website, will crawl it, will find information — but there are new ones now for the AI tools that exist out there. But anyway, sites that do a good job of organising themselves for those crawlers tended to do well.

Fan-out ranking. So last time, Mayna, you were on the show, we talked about query fan-out from Brighton SEO, which was the concept that every time you carry out a search on something like ChatGPT, behind the scenes there will be, you know, 20, 30, whatever different searches that happen on your behalf for related things. And what this research finds is that pieces of information that answer multiple facets of a fan-out query tended to do well. Which again makes perfect sense. What that means is that one-dimensional content is maybe going to be at a disadvantage compared with content that covers a topic in detail.

So none of this is rocket science, but it kind of validates a lot of existing thinking. And then there's stuff like matching search intent. So the example given in the article, I think, was that if the search term is something along the lines of "what are the best speakers to buy", list articles and stuff like that tended to do well, because they match the intent behind the search. Because asking "what's the best" sort of is a roundabout way of saying, well, give me a list of the top five or something. So yeah, none of this is groundbreaking stuff, but it's useful to see.

But the stuff that stood out for me that made this slightly different to normal SEO was that the organisation of content is important. It stood out that AI engines typically don't retrieve an entire page of content. They're more concerned with retrieving bits of information. So I think I might have said on a previous episode that GEO is potentially more about optimising ideas rather than optimising pages or content. And I think that's the fundamental difference now. You're sort of optimising to have a position on an idea, rather than saying, here's my webpage, I want that to rank number one. Instead, you want to rank number one for an idea or a statement, something like that. So that's the thing that stood out for me. But yeah, I don't know, Mayna, did you have any thoughts on this one?

Mayna (32:25)
I think this is another good example of reporting and measurement change, because a big thing from Brighton SEO is how rank as a metric is sort of becoming less and less reliable. And I think what we've been looking at, not instead, but alongside rank, is citation data. And Bing Webmaster Tools is a really good platform for that. It shows what sort of queries your website is being cited for, and you can use that to categorise them into sort of wider topics or ideas. And then it also shows you what specific pages are being cited.

Paul Wood (33:04)
Yeah, exactly. And I find that sort of information really interesting, because we've been using it for certain client sites, and using that data you can see — if you take like a few thousand search queries that are used in Copilot, which is where this data comes from — you can see where a site or a brand is kind of synonymous with a subject. So you're less concerned about ranking number one for XYZ keyword, and less concerned about traffic to a particular page, and more concerned about, well, we've got — in the last month or three months, we've seen an increase in the number of times we've been referenced against this topic or this subject. And that's really interesting, because then it tells you, okay, well, this brand is becoming an authority on this subject. And that's the end goal really, because then you start to sort of form a position in the market. And that's the bit that customers will notice. They'll start to know, okay, if I think about this subject, I think about this brand. And that's where the tracking becomes less specific, but more interesting to me.

Amelia (34:28)
So shall we move on to our final topic? And this is one — I doubt you've been able to speak to anyone this week without speaking about this particular topic, and that is the social media ban for under-16s in the UK. I mean, there's been a lot of conversations around this, finally confirmed by the Prime Minister this week. What are our kind of first thoughts on this, Paul?

Paul Wood (34:48)
Yeah, I mean, it's a hot potato, isn't it, this one? But I suppose my gut reaction is, it's a good thing. Something has to happen, that's my view. And I think the social media platforms could easily solve the challenges that the ban is trying to solve, but there's just no incentive for them to do so. I think a ban is a blunt instrument. I don't think it will actually solve the real problems, which are that they're addictive, and the content that you often are served by an algorithm is less than ideal.

The example I always jump to, the one I've seen happen in real time, is Twitter and how it's changed into X. I used to be a big fan of Twitter. It used to be a great source of news. If there was a big breaking news story, Twitter was so good for keeping up to date with what was going on. It was just really good for that sort of thing. And then as it transitioned into what it is now as X, you could see the change. Every time I logged in, the quality of the information I was seeing was poorer and poorer. It was riskier content, you know, really stuff you just didn't want to see. And it got to the point where I just left it for that reason.

And as it happens, there was a news story happening last week — I can't remember what it was — but I thought, it would be interesting to see if this is on Twitter, or on X, if anyone is covering it. So I actually created a new account, logged in to see if I could find anything. And immediately I was inundated with just awful stuff, like stuff that you just really didn't want to see. So again, I left it, because the quality was really poor. So if the social media companies aren't happy about these bans happening around the world, I think they've only got themselves to blame. But I can see why they do it, because obviously it works for them — if they keep people engrossed, they get the ad revenue from it.

Amelia (37:14)
But is that going to work moving forward? I think it'll be really interesting, because there's so many brands, isn't there? I think of beauty brands, for example, who — I mean, their whole strategy must be around aiming at teenage girls, for example. They do that largely, or probably completely, on social media. So where does that leave them moving forward?

Paul Wood (37:34)
Yeah, I don't know. And I think a ban, an outright ban, is a sort of — like I say, a blunt instrument. Ideally, there would be proper checks and balances in place so that these tools remain open, but the quality is better. But at the same time, I'm very aware that I'm not the ideal person to have a view on this, because growing up, social media wasn't around. I've used it in my professional life. Over the years I've become probably addicted to scrolling on certain apps. I'm a heavy user of YouTube, for instance, but I've got it to a position where it works for me, because I subscribe to stuff that I want to subscribe to, and yeah, it just works. But I guess, Mayna, you being a bit younger, you've probably grown up having it — it's been there the whole time. You've probably got a very different view on it.

Mayna (38:40)
Yeah, I've got a lot to say on this topic. I think it's very controversial, and I can see why. I think the nation's reaction kind of reflects my thought process as well. And I think the majority of people who are in favour of the ban are adults. And looking back at my childhood and time at school, as someone who grew up with social media, I'm more in favour of it now as an adult. But if this ban had come into force while I was still at school, I would have been very unhappy with it. And the majority of people I've seen who are unhappy with the ban are children under 16. And I agree with Paul, I don't think it addresses the core issue of social media being unsafe.

I also think the people who are going to be most affected by this are people who are under 16 who have grown up with social media, who will then have it cut off. And I mean, we discussed this yesterday, similar to the tobacco and cigarettes ban, how people born in a certain year will no longer be able to buy those products. I think rather than removing it from people who are used to it, and social media is a part of their day-to-day lives, I think cutting it off from a certain age group who don't already have exposure to that might be a better approach. But it's very hard to say, isn't it?

Amelia (40:08)
It is interesting, and you're saying it coming from the point of view of somebody who's perhaps under 16. And I think it's difficult as well, isn't it? Because there'll be lots of parents — I know obviously, Paul, you're a parent — who are terrified of social media when it comes to their kids and are really worried about the effect it can have on their kids, whilst probably sat on their own phone on Instagram whilst they're having these conversations, and children are seeing their parents. It's not like

Amelia (40:33)
they're going to be living in a smartphone-free house, is it? They're going to be seeing these things and they're going to be aware of them. And we know just how enticing and addictive it is. So it's difficult, isn't it? Are we ever going to get to a point where kids are — are we going to have a generation of children who don't even know what social media is? No, that's never going to happen, is it?

Paul Wood (40:44)
No. And I think it's so difficult, because it's a bit like alcohol, isn't it? Where parents have to sort of almost teach healthy habits with alcohol. A lot of parents will let their kids have a drink at Christmas when they're not old enough to drink out in a bar or whatever. And the whole idea of that is that you're exposing them to it in a healthy way. They get used to this idea of healthy habits, and you kind of want the same for this type of thing.

But at the same time, I think the companies behind some of these social platforms have really sort of taken the mickey. Because, for instance, X — coming back to X again — they often claim, well, we don't want to throttle free speech. You know, freedom of speech isn't the same as freedom to broadcast whatever you like. So I don't think that argument washes, really. Freedom of speech is about not being persecuted for expressing an opinion, but it's not about being able to broadcast to a mass audience anything you want to broadcast. And also, if you look at the history of Twitter of the past, they fairly successfully managed to keep people's streams of content fairly clean, free of anything too extreme. So they can't argue that it's too difficult to stay on top of this stuff, because I think they've got the technology to stay on top of the majority of the bad stuff that's out there.

Amelia (42:39)
That's actually what I was going to ask you, Paul, because I've heard a lot over the last few days, well, if these platforms had regulated better, we wouldn't be in this position. I always think — I don't know if you saw a documentary that Matt and Emma Willis did on Channel 4 a couple of years ago now, and it was about a social media ban. Well, smartphones, actually — they went into a secondary school and got a group of kids and made them give up their smartphones for a month. But as part of the experiment, Matt and Emma Willis set up TikTok accounts as, I think it was, 14-year-olds. They set up these accounts and they went through and they looked at content that typically you'd imagine a 14-year-old would go on, and they saw how long it took for highly inappropriate content — and it wasn't long, it was a matter of hours — and they were seeing content that was really, really frightening, that you wouldn't want your kids to see. And you think, how hard can it be to stop that happening?

Mayna (43:33)
I think it's a lot easier for companies to build on restrictions that are already in place than to have this open, free environment which anyone can use, and then put restrictions in place.

Amelia (43:48)
So it's sort of, yeah, it's too late now, essentially, what you're saying.

Mayna (43:53)
Here's a taste of freedom, but it won't last for long, you know.

Paul Wood (43:56)
Yeah. But I think this technology must be there to sort of manage a lot of this. But the natural incentive for these companies is to make the platforms appealing, or addictive, is the sort of way to maybe look at it. And you can't necessarily blame them for that, because that's the business model, and there's no regulation there to stop them from doing that. But I just feel like it needs something. It's a shame, because, like you say, there's a lot of good brands out there who will now suffer because they can't reach an audience with stuff that's actually useful and no problem at all.

One thing that took me by surprise when I was looking through the news about it was that it listed out some of the platforms that would be impacted, and YouTube was on the list. And I thought, I didn't even consider YouTube, because to me YouTube is just a really useful resource where I learn so much, it's just so useful. But thinking about it, that's because I carefully manage what's in my feed. I kind of

Amelia (45:17)
Yeah, I think with YouTube as well — I've got nieces and nephews of sort of primary school age, and they are so addicted to YouTube. It's wild. The things they watch, it's so weird. They watch these videos of people playing with toys. It's like you watch them playing, and they will watch it for — if you were to let them, they would watch it for hours, and they are like mesmerised by it. You can't even talk to them when they are watching YouTube. And I know that's something that my siblings — if we're babysitting them, it's like, you cannot watch YouTube for very long, you have to keep an eye on that. They're really concerned by YouTube, which is funny, because I feel like, as an adult user of social media, I'm definitely addicted to TikTok and Instagram, but I don't see YouTube as a real problem for me.

Mayna (46:00)
Yeah, we're definitely in the era of brain rot, as people call it. Like, I remember having a conversation with some of my friends about their TikTok feeds — and I'm sure there's things on my TikTok feed which aren't on theirs — but they were talking about how there's a series of Love Island on TikTok, but all the people are fruits. And they were really invested. They were so invested. They all had names like Strawberina and things like that. And I was just sat there like, what on earth are you watching?

Amelia (46:21)
I've seen this. I know, but it is like that, isn't it? You watch these things and you sort of get halfway through a video and you think, what on earth has this become? I got really addicted to this series of just a guy who cleans up old bits of makeup. He gets old bits of half-used makeup and he cleans it and restores it. And I'm like, oh my God, I'm 15 minutes deep into this, what is happening? But you know, we can see how it happens. It is so addictive. But I'm sure, again, this is a topic that we will come back to.

Amelia (46:57)
So before we go, should we have a quick round of jargon busters?

Paul Wood (47:01)
Yeah, let's give it a go.

Mayna (47:02)
I'm up.

Amelia (47:03)
So this is where I bring you an industry term to see, quite simply, if you two know what it means. And today's term is beta.

Paul Wood (47:12)
So is that — that's essentially something that's, you know, in testing, isn't it?

Amelia (47:21)
Possibly. So beta, or I think you can pronounce it "bayta", that's the American way of pronouncing it. Mayna, have you got any thoughts?

Mayna (47:30)
I'd say the same as Paul — something in trial are my thoughts.

Paul Wood (47:34)
Yeah, because in the world of technology, when you launch something in beta, it's a sort of test version. You'll often see apps and things that have the little beta icon. And that just signals to users that this is an early access version, so don't be upset if there's an error there, because we're still fixing it.

Paul Wood (48:02)
But in the world of finance, yeah, I'm not sure if it will have a different meaning. Is it a test sort of, like a test fund or something?

Amelia (48:16)
Shall I let you know? Kind of on the right line. So it's a type of measurement of risk. So it's a measure of a stock's volatility in relation to the overall market.

Amelia (48:29)
So if you're testing that risk, it's kind of along the same lines.

Paul Wood (48:33)
Yeah, but it's quite different to the world of technology. Okay, that's an interesting one.

Amelia (48:41)
Yeah, another one to add to the list.

Paul Wood (48:44)
Yeah, every day's a school day.

Mayna (48:46)
Yeah, learn something new.

Amelia (48:46)
Exactly, exactly. And of course, we will have more jargon busters next week, and we will be back next week. Brilliant chat today, guys. Thank you so much, Paul and Mayna, and we will chat with you guys soon, and we will see you next week.