
There is there is a good garden. Thank you. To start back. Thank you very much. Of this baby on and your finger. Play that recording on it. Like the first. Yeah, yeah, yeah. You can play them. Just have to send it to the right file and say it back.

This. Okay, I'm gonna start there. He coming to the ground bruise, and yeah, so yeah. So just giving you some ground rules, they're probably the kind of thing you'd expect. So we would encourage you to participate. Actively, Let us know what you thinking. Try not to kind of talk over each other for the sake of the recording.

I'm, although I can be guilty of it myself. So don't worry. Another tell each other off to yourself. Yeah, exactly. We're human. Um be respectful of obviously please and for focusing on the topic be open minded so that you accept other people's viewpoints and maybe that will help you think about yours.

Um we're going to maintain some confidentiality here. We're not telling anybody else. What has been said in the room and who's said it Um, we are hoping for some honesty transparency, there is no right or wrong because it is really about how you felt about what you've just done and we are recording and taking notes.

Thank you. Ice cream. So the beginning with the focus group, the first question is about your views on misinformation So, the question is, what? Might be the effect of mission information. What do you think might be the effective missions information? Obviously, people just getting the wrong data in facts and That just leading to them.

Maybe doing the wrong thing, you know like anti-vax people or whatever. Yeah. Give us spreading third misinformation. That's right, wasn't not. Yeah, spreading it. And anything else? I think there's, there's Bit of a bit of danger when Like the misinformation. Is. Contains to something, which affects people's reliables, or, you know, things like that.

Where, you know? Where something that's that's bad while. In some cases, it may not affect the outcome all that much, but in other cases, it can drastically change. Yeah, I think there's levels to the misinformation of all level. Sure. How, how would you like when you say I was just lying to him about which direction to go compared to if you know the media or whatever was saying That happened.

But in reality it, something else happened, you know, of course, to very different things. Effective. Thank you. So, on the next set of questions are about Of your experiences with misinformation. Right. So the first question there is, how do you check the accuracy of misinformation? But try and look up.

The first person was expected. It doesn't information or just look up. See? Welcome fine. Or if it's If the wording of sadness, Information is bias or try and reword it in a non-by-square and look up and then just see different new points. And, you know, just make up my mind from that.

Right? Interesting. So you'll take the same like phrasing. You're just trying to get rid of what you think, advice, terms, try and generate a search. Yellow. Look for example. For someone, go on a run about a particular statistic. Okay, can I just Okay, so like the 13% is that And the fuel familiar.

No, go ahead. Tell us a slightly. Is it first himself? It's like, 30% of black people, commit 50% of the crime, Right? Right. So like people share, that's cystic everywhere. For like they won't look into it. Like they don't even know if that's this complies That's why they live there and take you to talk about proximity.

He's It's like the way someone can work that and share it. It's blazing misinformation. Like, And, Yeah, like something like that. It's harder to research. It's like makeup your own mind because a lot of people are just not going to read studies. They're not going to read. Yeah. It does require it.

Yeah. Get down to like, looking for the survey findings or whatever. Like track probably. Yeah. Pull to get that information. It's really good point. Yeah, i wasn't aware of the 1350. So yeah, that's that's like one of the Biggest piece of misinformation. That's like shared around. I'm not going around.

Yes, years of it years of life. Position on social media or is it? I mean, yeah. Social media. Maybe like only internet. Yeah. Of. It's living meme. No. But he's still here the prejudice, right? But I wasn't aware of like the numbers had been assigned, right? Because you needed and passes.

I wasn't aware that. Yeah, yeah, more time. You spent online, you find it right? Eventually find the eventually. I think man, the worst places. Well, do you know where I mean? It's that other particular size. Was it literally like it's just me social media, You TikTok and Instagram. I mean the same people that share this, they just show everyone, they can't.

Yeah, yeah. It was that? No. Just oh, that was some juicy information. We just Okay. Up some moving on. Are there any other ways of checking the accuracy of national information of information? I mean considered source right is is quite a big part of what I doing anyway is so if someone has said something on Twitter I'm never going to take that a face value.

You know? It's It's it's you know, if if And people say things on social media, especially if it's like independent people who don't. There's no, I mean, I don't know if there's regulation anyway, but I assume there's some regulation for news outlets. Um, the determine how far they can bend the truth.

But a particular for individuals, there's no regulation for that. So you can't really trust anyone, but sure. I'm more likely to trust things from news outlets. But like, I would take Like a broad spectrum on things that I know work you know to the right and then to the left and into the middle and hopefully trying to average it and try and find what the common ground is between.

Food. Yeah, i'm so how do you distinguish between the two? So you say sources or something? You look at. So are there specific sources that you think? Might have misinformation. I think probably i generally speaking, I reckon all sources are probably got some elements of misinformation to them. But, And you kind of have to go.

Okay, so these ones are going to stretch this bit of information. And these ones gonna stretch this in a bit of information. And you know, in the case of independent people they can make up whatever, right? Um, So it's the question of trying to find a diverse enough range of sources and then trying to Somehow find out what the common thing is between all of them.

And take that as that way, you go for your kind of ground truth I guess. Yeah it's the common. The commonalities that run through most likely any you sipped out you go. Okay well that only that only appeared in one of them. Yeah that might not be so true And okay.

Yeah, that's something that you guys also do to you. You look at everything. One of those look anything on the news, like, I don't take all my information from just one so the new source that's a new source. I might look at another new source or individuals online, but maybe I follow and see what they think about it.

If you, when you're looking at news sources, do you look at new sources that are UK based? Do you look at other countries than years? It's all. I guess it depends on the news but usually I just whatever shows are you know what? For example, this something we would be Israel colour science or computation like I'm not just gonna read what CNNs has about to.

Yeah. Yeah. Yeah. Yeah, so yeah yeah. I mean that my I have a relative who look at what what Russia TV would say. Verses I'll just zero versus the BBC. London news every morning just to kind of get a sense of like okay. Well they're lying about that, they're lying about that in a bit like yeah.

What's the what's the common true? Not really interested in. Yeah to do. Yeah, it makes it on the same topic. It brings us to the next question. What kind of information do you think is most prone to being misinformation? Because this looks, Yeah, this is everyone is like an arms.

The armch after station nowadays? Yeah, no. Like you you can have as much experience you cannabas as much knowledge on a subject as you want. I'm sorry. I read this statistic online, three years ago and not prompts everything you're about to tell me. Yeah, but yeah, I think there's because The most easy pool supply will things on taking off contacts And Absolutely.

The person's card. First is the thing that takes Requested over any reasoning that I'm usually nation or is it about the numbers of? So it's mainly statistic because people can hear like people, see, things see storied like okay, well he's he said that someone raises a statistic. Okay, this is fact right?

Signals from being an anecdote. Yeah, one person tell you story to being like, well, there's numbers behind this. Exactly. I thought like it depends on like the kind of person that tells you as well, okay? Because if like somebody who just Around a lot, gives me a number. I'd be like.

Are you sure about that though? It's like they've got gravitas or you trust them to be like yeah so they just told the story really well I guess. Yeah. Yeah so it the source matters. So yeah. So are there any specific sources? Which you would Rely on or wooden rely on.

Okay. But I'll say I've been online enough to not rely on any sort of yeah. Yeah, no trust. Anything that are sources that? I trust slightly more than others but overall, I do not trust anything online. If you like, you've got to take each information that's own little thing And now some reading like something on Google Scholar.

Yeah. On I'm not taking any opinions or any anything from people online, right? I mean, I think, I think there are probably some new sources for certain things that are more trustworthy now, it's, for example, I would probably take what the BBC says over. What the, you know, Russian state has come out and said about Ukraine, but about Ukraine.

Yeah, you wouldn't you wouldn't necessarily The biggest says is about but you wouldn't, you certainly wouldn't, you know, take what Russia says those were made and for that sort of thing. Um, because they're extremely likely to focify and something. So that it down to also the context. So, what's the subject as to whether or not when you make in your comparison system who's likely to be more trust with the, it's going to come down to also the topic not just the sources.

Yes, yeah. Okay. That it's now. Ohm Moving on to the next subject, thats your experience of using the tool. Um, to what aspects of the tool? Did you find easier difficult? So other specific things that you found easy or certain things? That would more difficult. Because it's very easy to navigate the findings of what that's all picked up.

Hmm. Okay, so that difficult thing is this like the system, the use for ranking, or like sometimes sender, just sometimes it's fractions. Sometimes they are negative like From like I'm not sure what the units already, or what the range is like calculating me as well. How is calculate here?

I hope. Mainly, I do think the the right, the range comment is an extremely important one because The like I didn't know this until that one score went minus that it could go that way and I was like, oh okay. I assume that was between zero and more but now it's gone mine.

So that's like yeah, yeah. Now that's skewed me. Yeah. Yes, it's something about. Is it something about the fact that they're all been different? You don't know what they mean. Would you like to see the scores? Either explained, or which rather have uniformity in the way that they're presented.

What would what would make the boat for it? Would be nice if you want to uniformises but then in a tour like this, we're trying to have Of like the The most users, I think it's just a uniform zero to 100 of each. Okay, yeah, that works and also good of context on what they're actually murdering because, for example, a lot of polarity one, Someone might be reading an article, They don't know what the polls are.

So if it's like polar other, let's come back for something that like I don't know like what that even yeah, it's referring to so some context Control the things that just picking up sort of clear subjectivity is the things. Yeah. So we want some different, some definitions for the terms and some uniformity in the way this scored so that we try and stick the same scale or the same thing numbering system or something like that.

Yeah I feel like that would be better but like If that's not possible, then I think the next best thing is to show the range somehow, so you can tell where it folders in the range. Yeah. Okay. I think they look at the sliding skill. Yeah, something good though.

Just a number. Yeah. Yes. And I just show you number show you where it says, yeah. Yeah, I agree. Actually it could help me anybody. Me some kind of overall school. Might be nice as well, like, for the entire article, Okay? And can you ever got any idea of what that might look like?

What colour? Just, if you add up, like, all the variables together and then give an average, yeah. Given an average and say, like This 50% light need to be Okay. Yeah, so an average of all this voice, Take a note that German Good stuff. Tom Moving on to the next question.

Or do you have any concerns about using the store?

You don't really again. I mean, they take It works. I mean, if it picks up everything, I feel like even with this, I still end up going to another website. Checking that audio. Formulate. Yeah. Okay. So would you use this tool as a basis for your product search? Yeah, I think probably an explanation of how the tool.

Of figures out, what it, what it did, what it means to be. You know, that this is Of what the scores. How it figures the mouth because And you can you can thing is if you make it a black ball box and it's it's all You know, all self-contained and you don't really understand what's going on inside it.

Then. Of people are less likely to trust it because it's effectively just another new source. Know that point if you explain to them what it does and you know it takes this thing and compares it against all these things and It's looking for these particular things and this is more of this that in the other.

Then you end up with people going, oh i you know, I understand more what this is actually doing. So we're more likely to then trust. It's an analysis of the Topic, or whatever. Fair enough. So, like if you had all that up front. Yeah, yeah, good explanations. And is it just about having a kind of like an instruction manual before you know, or use the guide up front or is there something about keep this?

Do you think the uses gonna need reminding of things as they're going through it? You know, is they're going to be a need for more products. Of the night. You know, when you like hover over something I could description comes over. Yeah, something like that. Might be like that.

So like just a cover you in case you kind of forgot or you're not sure because me again. Yeah because even if like if it's one person using that laptop somebody else could come along and use that out of anyway. And just not know what's happening. Thing maybe like, One of the question mark icons top right?

Teaches all web page that explains more about the tone out works like us. Okay? Because that's that's not quite a lot of extensions were. That's a little bit. Little question mark. Yeah it gives you what you need and take you to a healthy and tutorial. Yeah, it's great idea and the university little simple, isn't it real?

And look at Anything. Um, Bring that so the Trust brings us to the next question. Did you trust the store to flag language often foundation information? But trusted the flag language within us like in misinformation. I don't trust it to see misinformation. Yeah, I think it needs more of a And contacts, like some way to figure out the context.

The two other any specific factors that would help you trust. Little more than you don't know, I think Just making it overall because more accurate because there are some things where it identifies it and that's, you know, it's the result of a study. So it's not misinformation per layer.

Like some rules might be nice because there was a One of the variables. It's my first person language, right? And for I think they don't want to save corn or whatever and it was just quoted from him. I thought like that doesn't need to be flagged because it's a quote, rather than You know, the right to being like, I could even this and did it.

Well, you should be like Should be very, very careful about presentative forces. Because, In a tool where you're using it, your using this to see short, trust us information or not. And you start reading false positives, that's going to feel like a great deal of distrust and not too.

Even I think, oh well, this is just readings. I'm going to trust them. Just gonna like read this for face value. Yeah, false positives are real. So I think like, just dial forcefuls down as far as you can. Risking removing some things that are actual like misinformation. Yeah, just even did not have so many false positives.

Yeah, good point. Yeah, because that bit in the accuracy. Um, calculation is, is problematic for people? Yeah, I think you're right, we just makes you think. Don't think. Um, What do you think might be the F effect of accuracy on trusting a tool? Actors. So seconds, massive Yeah. As long as it presents me that actually score.

It says, okay, or flag this or about. The tempus. I'm sure that this is misinformation As long as it's as long as it tells me that then I like okay you happen to see a number yeah, no matter what number is you want to see a percentage and you want to see a number of scale or some kind of thing, What like would make you happy that you want to see the number see you know what it thinks but what kind of number would make you comfortable that this is probably misinformation or things thinking more just define those numbers inside the program and ranges and just run the results.

Based on like, how much it trust that's misinformation. So like I click on subjectivity and the top thing I see is that then with the highest accuracy, And it feeds down, and you have maybe like Them all over, say traffic like You know like this top one's like it shows, they're just indicates somehow the top ones have a higher weight in the low once.

You know, and when you say the top ones, the top ones is being the most accurate or the most. Yeah, inaccurate. Most popular most like just see the most, the look, the most accurate things that it knows I miss information on first because, you know, Because you don't want to see like 43% of 43 for something.

So and then there's like a 199% right underneath in this just someone's not more than just scroll down read. Yeah, yeah. So yeah, you want the big things to stand out at the top. I think there was there was an interesting point in that And that was the Right now, i i don't know whether the percentage score that is giving me is an indicator of whether it's 50%.

Sure that some of it is information. Yeah. Whether 50% of it is, instance, information. Nice, you know. Okay, yeah. So the huge difference between that's, you know, that's yeah. So my 50% sure that there's misinformation here or is 50% of this misinformation. Yeah. Okay. Yeah. We need to make that a lot of clear, I think it'll be good if you like to just clearly stay at the The pool doesn't actually check misinformation.

It misinformation is just looking for languages as well, That was absolutely right. And that's the sticky point where it's so easy to just go. Oh, this is the tool that identifies misinformation but it doesn't doesn't it identifies works that maybe. Yeah. Out here. And as you say, it needs to kind of that needs to probably people need to be reminded of that because even way sometimes make the mistake of saying that it's a tool that identifies this information and no no no.

It's easy to slip into that. And yeah, really is human heuristics. We just fall into it like that, 95% of, you know, I guess not even, right. Any more questions? Um, yes. So thinking about the information provided by the tool Would you want the information to be provided in a different way?

I think in some cases it would be nice to have because in I know some cases it does highlight the words that it thinks Michaels but in some other cases it's gone like here's a block of text and I've gone what in that is. What you might not find and, you know, it's it's not playable to So I think Pointing out specifically what it is that it's identified just.

Yeah. Because in context it might mean something completely different. Yeah. You know. Good point And anything else you would like to aren't? I think that's just a very good point. Like, for some of the texts like subjectivity flag. The entire article, she's I don't know. It. I love you bullet points each book.

A paragraph and a half, so yeah. It's a big man. Maybe we should give it some more subtle articles. Yeah. Yeah. The whole particular program that article might have been an outlier, I hope anywhere. I hope that it didn't do that with everything which is like highlighter To make your day.

Know, it's a good point. So yeah, picking out the actual specific bits and not being so. Scoop. Yeah. Um, is there something? So it's so what information would you want? That is not provided in the tool already. So, the variable stuff, obviously, Other than that, will just inflammation on the tools.

So, yeah. I think that's been so tired. Yeah. But, He's just a bit more, but the very simple, very full stuff. Yeah, just like defining what subjectivity means. So, yeah, okay, yes, yes definitions. Me. And we mentioned racial ranges already as well. I think it's yeah to have a range count with that's a definitions and arrange just Here, the overall school stuff, I guess.

Both of you. My last shit. Of just, I just wanted to go back to the effects of this information. Team, who do you think is either? Of the most susceptible or vulnerable to me information. It's slightly or who do you thinks the United States down by being exercisity. You can look at a lot of smiled I hope we're all thinking.

It's going to say before people. Yeah, yeah. But the unfamiliar with yeah, internet like yeah. It's because for us, the look. A grown up on the air. We've like, we've We got that with like with nothing important. To our name of that makes any sense. Like we don't have like bank information that can be scammed up with, we didn't have liked things at that So we feel for all the traps but that risking anything but with people that's like them just gone into the internet.

They actually have lives to risk and life. House, right? Because the empty. Exactly. And we hear those stories as well, Like people that they can influence with misinformation, things like that. Yeah, I think that they're not that witty upon, you know, they're not that up on the internet. They can be kind.

They can pass on and misinformation to other people. Yeah. Because they'll fall for it. But also like I feel overly trusting people who just, you know, they're read something and go oh yes, these people have said it. So you know, they're a big corporations so it must be true and then you know, they don't go and check it and then they start spouting to other people who go.

Yeah, you talk you back But then they get into one doing some access it. So and do you think that's come down? Sometimes the sauce like people trusting a source? Yes. And then thinking, well, you know, that got to be the legitimate because it's come out. Do you feel something like that?

And then somebody else would be like, Yeah, yeah. So that's a source issue. It's about trusting problems isn't today. It comes to hold popular as anything. It wasn't it. But before characteristic feeded, you've got some Donald Trump forest jokes and not just, they're all really popular leaders and people trust them.

You know? They're they're mates their friends that got a tone they speak. Do you like, they know, it's really good. They use their charisma. Yeah. I think close. Are we we? Yes, that's the end. We have any other questions, Any questions, anything you have that? You'd like to add any other questions?

Anything, we have enough that you thinking about mulling over Alive. There isn't it right? Because you go, So much, think about, I think Jeremy is got something for you. Yeah. Then we let you go. Yeah. Thank you so much. Thank you dressed up these. Up like this.