This is the AI service you should never sign up to

**Lianne Potter** (0:03)
Name 10 punctuation marks. 10? 10

**Graham Cluley** (0:07)
Well, you have an exclamation point, you have a colon, you have a semicolon, you have an M dash, you have a hyphen, you have a Spanish upside down exclamation mark, you have a comma, you have a full stop, you have a slash, a colon, have I said?

**Lianne Potter** (0:27)
You've got two more to go, two more to go.

**Graham Cluley** (0:40)
Smashing Security, episode 480 This is the AI service you should never sign up to, with Graham Cluley and special guest Lianne Potter.
Hello, hello, and welcome to Smashing Security episode 480 My name is Graham Cluley.

**Lianne Potter** (0:55)
And I'm Lianne Potter.

**Graham Cluley** (0:56)
Lianne, welcome back to the show. Lovely to have you here again. For those people who are not familiar with your work, what do you do and what have you been up to lately?

**Lianne Potter** (1:06)
Well, it has been a long time, so I'm going to hold you very responsible if I scream. Oh dear. For those of you who don't know me, my name is Lianne Potter. I am currently the Chief AI Security and Ethics Officer at North Star Intelligence, which is a new gig for me since I last was on this show. I've also got another new podcast which finally came about. I know we discussed that on my last episode, but it's here.

**Graham Cluley** (1:31)
What's the name of your new podcast? Because you were doing Compromising Positions, weren't you? What's the new one?

**Lianne Potter** (1:36)
Compromising Positions is still going great guns, so it's there for you to enjoy, which is my Cyber Security podcast. Not the same as this one, but Variety is the Spice of Life, right? Yeah. But the new podcast is a lot more lighthearted. It's called Tech Film Noir and on it, me and two other techies review the technology in science fiction films to see if it predicted the future.
The latest episode, which will be out when this episode airs, is For Strange Days, which is a 1995 classic, in my opinion, that no one has ever heard of. It's all about VR and living over people's memories and stuff. It's quite dark, very cyberpunk, but very fun.

**Graham Cluley** (2:16)
I found it quite disturbing, I think.

**Lianne Potter** (2:18)
It's really disturbing, yeah.

**Graham Cluley** (2:20)
And you've been doing something with incels lately. Very good of you, very charitable.

**Lianne Potter** (2:25)
I mean, as a woman, yeah, I'm not sure I'm the type of person they want to hang out with.
I am currently finishing off my masters in AI. For my dissertation project, I have been building synthetic incels. The reason being is because I just can't get hold of any real ones. Incel stands for involuntary celibate. It's a subgroup of the manosphere that believe that they're capable of having romantic or sexual relationships because society is against them.
I've been thinking a lot about, well, how do we de-escalate this de-radicalization because it's all connected with how they engage online. It's a very online community. Incels are an issue because they come with a lot of negative violence and danger.
Compared to the rest of the male population, which has a suicide rate of about 3%, incels tend to have a suicide rate between 43% and 62%. Oh my goodness. Yeah. There's a high rate of mental health issues within that community. They're also seeing a lot of incels use things like AI companions, AI girlfriends, to enact violent tendencies. So I thought if incels are using AI companions to practice and rehearse negative interactions with women, how about we get that AI companion to challenge these beliefs? So I'm doing a thing where this AI companion, if it detects this incel is starting to veer off into radicalization, it will then start challenging those beliefs. I've done four or five different experiment tropes.
One version is it uses persuasion techniques to try and convince them to maybe not think like that. Another is a Socratic prompting. So getting people to reflect back on the reason why they think that. Various other means of behavioral science to understand what would be the best approach if we were to apply this AI companion in the real world. The results are that it does slowly deradicalize these incels. Now, you're probably thinking, how do you make an incel? Well, unfortunately, I had to troll through absolutely loads, and there is a lot of data sets out there, of 4chan, Reddit sub posts from these fellas. Not the nicest discussions to read, but all made a very convincing incel. So that's what I've been up to.