Neil Kirk

Too Authentic to be Synthetic: The Psychology Behind AI Voice Scams As AI voices become eerily realistic, new research shows that the instinct to trust our own accents could make us more vulnerable to deepfakes

As synthetic voices become indistinguishable from real ones, DIGIT spoke with psychologist Dr Neil Kirk about how our instinct to trust our own accents could make way for deepfakes targeting not just individuals, but entire regions.

www.digit.fyi/psychology-b...
#deepfakes #AIvoices #AyeRobot

Too Authentic to be Synthetic: The Psychology Behind AI Voice Scams As AI voices become eerily realistic, new research shows that the instinct to trust our own accents could make us more vulnerable to deepfakes

Here's a feature on my recent AI Voice work - really honoured to have been asked to talk about this! www.digit.fyi/psychology-b...

🧠 Is your mindset making you more vulnerable to AI voice-based deception?

I've got a new pre-print out: osf.io/preprints/ps...

Vigilance towards AI voices can be nudged through a change in MINDSET

Thread below 👇

#AI #Voice #Cybersecurity #Fraud #Psychology 1/11

Very grateful to @the-sipr.bsky.social for funding this important work. 11/11

💡 Why it matters: This could have real-world implications for designing public awareness campaigns and scam prevention messages. 10/11

🏠 Take-Home Message: Simply telling people that AI voices can speak with a Scottish accent/dialect was far more effective than warning them to be vigilant. 9/11

However, an explicit vigilance-based nudge warning about the dangers of AI voices and urging listeners “if in doubt, think AI” had no effect, unless paired with the capability message about AI’s linguistic abilities. 8/11

A positively framed nudge highlighting AI’s capability to reproduce underrepresented accents and dialects significantly reduced this bias – in other words, changing their MINDSET made them more vigilant towards AI voices using these varieties. 7/11

In this manuscript, I investigate whether simple informational nudges can shift these assumptions and reduce the bias for responding “Human”. Across two experiments, participants categorised voices as either Human or AI. 6/11

Yet that assumption could be putting some language communities at greater risk of AI voice-based deception if they believe a voice speaking that way must be a real person. 5/11

In my new paper, I introduce the concept of MINDSET: Minority, Indigenous, Non-standard, and Dialect-Shaped Expectations of Technology. It reflects the idea that people assume AI can’t convincingly reproduce underrepresented ways of speaking. 4/11

I also suspect this is not unique to Scotland, but part of a global pattern affecting communities whose voices have historically been excluded from these systems. 3/11

My previous work showed that listeners were more likely to believe an AI voice was a real human when it spoke in a local dialect. I think this happens because we’re not used to speech technology understanding these varieties - never mind speaking them! 2/11

🧠 Is your mindset making you more vulnerable to AI voice-based deception?

I've got a new pre-print out: osf.io/preprints/ps...

Vigilance towards AI voices can be nudged through a change in MINDSET

Thread below 👇

#AI #Voice #Cybersecurity #Fraud #Psychology 1/11

New achievement unlocked - delivering the oration for Brian Cox. I didn’t faint, nor did he use any of his famous catchphrases on me. Success!

<em>British Journal of Psychology</em> | Wiley Online Library Self-cues such as one's own name or face attract attention, reflecting a bias for stimuli connected to self to be prioritized in cognition. Recent evidence suggests that even external voices can elic...

Read the paper here 👇🏻 bpspsychub.onlinelibrary.wiley.com/doi/10.1111/...

Made another wee video about one of our recently published papers…

… or take Monday off. (Yes that was me).

I’m on TikTok and Instagram @ drspeakmind - please give me a follow and see how well you can spot AI voices! 🗣️

Three-panel composite image of a man completing a muddy obstacle course: left shows him gripping gymnastic rings mid-air, center captures him grimacing while emerging from muddy water beneath a black pipe, and right shows him smiling and soaked after finishing an icy water-based challenge.

“Enjoying” some time off work #ToughMudder

🎓 Fully funded PhD at Abertay University & SIPR: Enabling People to Identify Deepfakes. £20,780/year, fees paid, starts Oct 2025. Use eye-tracking to research deepfake detection. Apply by 30 June 👉 www.abertay.ac.uk/about/workin... #PhD #Cybersecurity #Deepfakes

Me holding my Nintendo Switch 2 console

It’s always lovely to receive a “to me, love from me, well done on finishing your marking” present.

Mr Predator and Mrs Prey: Gender Stereotypes in Children's Films Correlate With Explicit and Implicit Gender Stereotyping Children acquire gender stereotypes at a young age and these subsequently influence cognition and behavior. Stereotypes may be learned through a child's direct observation of gender differences as we....

🚨New publication... Children's films contain innacurate gender stereotypes and these corresespond to children's (and adult's) implicit and explicit gendered associations.

onlinelibrary.wiley.com/doi/10.1111/...

Read thread for more details...

This might make some language communities more vulnerable to AI voice-based deception. Luckily I’ve been given some funding to investigate this further, so watch this space!

Do they factor self-indulgent puntastic titles into REF scores? I sure hope so! 😄

This might make some language communities more vulnerable to AI voice-based deception. Luckily I’ve been given some funding to investigate this further, so watch this space!

But this bias was stronger in a regional dialect for those who were more familiar with it.

I think this is because prior experiences with voice technology that doesn’t represent (or understand) regional or minority language varieties made listeners think “AI can’t speak like that”…

In two experiments I showed that listeners had a bias towards responding “human” when asked to identify if a voice was a natural human speaker or had been augmented by AI to change the voice into a new vocal identity…

“Eh? Aye!”: Categorisation bias for natural human vs AI-augmented voices is influenced by dialect Advances in AI-assisted voice technology have made it easier to clone or disguise voices, creating a wide range of synthetic voices using different ac…

I’m extremely chuffed my new paper “Eh? Aye!”: Categorisation bias for natural human vs AI-augmented voices is influenced by dialect” has just been published in Computers in Human Behavior: Artifical Humans.

(and not just because my best work is in the title!)

www.sciencedirect.com/science/arti...

Never knew hearing a Drag Queen play a medley of 90s dance hits on the saxophone could be so healing 🧘‍♂️