Last week a global consultancy firm hosted the first in a new series of client webinars and invited our Director and former Head of the Police Fraud Squad, David Clarke, to share insights on AI assisted fraud and the threat of deepfakes.
Using AI generated videos, Mr. Clarke, a leading voice in counter-fraud and secure translation software development, explained how fraudsters are abusing AI on an industrial scale, but more importantly, how secure, safe AI is the essential weapon needed to defend against it. He framed the biggest challenge today as a battle between malicious AI and trustworthy AI, and it is up to us to wield it as a tool to protect.
Download his presentation here:
Tracing the chilling ‘Evolution of Fraud: From Faxes to Deepfakes’, he showcased how scams have progressed from traditional ‘419 letters’ to automated, hyper-personalised, and highly scalable digital attacks. These are visible to all through the increasing volumes of spam emails generated by AI agents that scrape websites to find information to personalise messages.
A key focus was the corporate threat, highlighted by the shocking £20 million loss suffered by the UK engineering firm Arup, where an employee was deceived by an AI-generated video call that mimicked real colleagues.
As Mr. Clarke stressed, the ease of creating ultra-realistic digital human forgeries - known as deepfakes, demonstrated using Grok to make a video of President Donald Trump speaking about his love for Luxembourg, means the digital world is increasingly difficult to trust.
Pivoting to the positive, he then showed the protective power of AI, with global organisations that use Guildhawk avatars to present safety information to people in over 100 languages powered by human verified multilingual datasets, sitting on a secure private cloud.
There is a significant between deepfakes used by bad actors that are not restrained by the law, GDPR and ethics, and corporate digital humans such as those created by Guildhawk. Deepfakes are deceptive, non-consensual forgeries designed to trick; Guildhawk's Digital Humans are ethical, authorised digital replicas, designed for professional, transparent, and scalable communication, such as multilingual safety training or compliance videos.
"Our digital avatars are created by professional personnel and stored in secure vaults, and all your information and data are protected by our ISO-27001 standards. "
explained Mr. Clarke in his presentation.
The core of Guildhawk's defence strategy lies in our proprietary technology, the GAI small language models (SLM). Mr Clarke told how SLMs are the future of anti-money laundering (AML) and fraud detection, offering a significant advantage over generic large language models (LLMs):
Mr. Clarke’s keynote concluded with a powerful call to action:
To protect against the escalating AI threat, organisations must prioritise data privacy, staff training, and the adoption of secure, domain-specific AI solutions that turn the tables on the bad actors.
Guildhawk is proud to be leading this charge, delivering linguist-level precision and ironclad security through our GAI SLM technology to help global businesses safeguard their people, data, and capital in the AI era.