In July 2024, a Ferrari executive was targeted by deepfake scammers. Learning how to spot a deepfake is crucial as they employed a voice clone to imitate Ferrari CEO Benedetto Vigna, following several WhatsApp messages.
However, after some red flags were raised and a transaction was requested, the executive asked a verification question that only the CEO would know. This caught the fraudster out and put an end to the deepfake attempt.
Trustpair’s fraud prevention software helps prevent financial losses that occur from cyber fraud. Our advanced algorithm automatically verifies third-party bank details and transactions in real time before they are executed to offer added security.
What is a deepfake?
A deepfake is a digitally manipulated piece of media designed to appear authentic. Using artificial intelligence, deepfakes can create convincing videos of individuals saying things they never said, being in places they never visited, or even fabricating events entirely—making them a powerful tool for spreading misinformation or fake news.
However, in some cases, artificial content is indistinguishable from real media.
The process starts by feeding deep learning software pictures, videos, and voice recordings. The software then absorbs and learns this information, and it can be added to other material to make it appear real.
The usage of deepfake technology can relate to face-swapping applications and to age or de-age actors. It can also have a more sinister use as deepfakes can be used to commit blackmail, identity theft, and fraud to businesses.
What is the impact of a deepfake on businesses?
Financial loss
If a fraudster manages to commit deepfake fraud in a business, they may steal thousands or millions of dollars.
Companies must have measures in place to avoid the risks of deepfake fraud. After all, 20% more companies have reported deepfake video incidents in 2024 compared to 2022. Also, audio deepfakes have risen 12% in the past two years, with nearly half (49%) of companies facing the threat.
Platforms like Trustpair help protect companies from financial losses caused by cyber fraud. Our account validation process ensures the company is legitimate, verifies the bank account’s existence, and confirms the match between the two.
The account validation process typically takes 30 seconds, providing companies with quick assurance against cyber fraud.
Reputational loss
A deepfake can significantly damage your company’s reputation, potentially leading to a decline in stock value as public trust erodes. This loss of credibility may drive customers away, directly impacting revenue streams. Additionally, stakeholders’ confidence in your organization could waver, creating long-term challenges in securing investments, partnerships, and market stability.
For example, if a deepfake video positioned a CEO as saying something discriminatory or false, this could impact and influence the way people view the business and whether they want to work with you as a buyer. It may also mean that potential employees think again about applying for a job at the organization.
Take this example from The Guardian earlier in 2024: Mark Read, the CEO of WPP—a large advertising group—was targeted by a deepfake in a CEO fraud attempt.
Cybercriminals set up a WhatsApp account pretending to be Read and created a Microsoft Teams meeting that was said to involve him and a senior WPP executive.
The fraudsters used a voice clone and YouTube footage of the executive. The unsuccessful scam was thought to target an ‘agency leader’ to get them to start a new business to obtain money and personal details.
Had this scam attempt been successful, stakeholders’ trust in the senior executive may have been damaged.
Tools and technologies to help spot a deepfake
Several tools and technologies can help identify deepfakes and combat fraud. While not 100% foolproof, these tools can provide valuable indicators for detecting manipulated media.
- For images – Illuminarty detects AI-generated images, manipulated images, and deepfakes. Google also provides Reverse Image Search so you can see where images may have also appeared
- For videos – The scanner tool Deepware detects manipulated videos
- For images and video – Hive’s deepfake detection model offers a classification and a confidence score for the detection on a scale of 0 to 1 for the content.
- For images, videos, and audio – Sensity (a paid tool) analyzes visual and contextual cues to detect AI-generated content.
How to train your team to recognize manipulated media
Media manipulation involves altering a piece of online content (audio, image, video) to reach a goal. This goal could be to damage another company’s reputation or encourage customers to buy your product.
There are several tips you can give your team to pay attention to so that people can notice deepfakes more regularly:
- If the audio doesn’t match the video – for example, the person’s mouth may not be saying the words that are coming from the audio, or it may be too quick or slow at times
- Robotic pronunciation – if a word sounds like it has been said like a robot, and the person ‘saying’ it wouldn’t usually sound like that, there’s a chance it is a deepfake
- Odd movement (usually in the head or body) – if they move unnaturally in a video then this should raise suspicions
- If the background is confusing and makes little sense – for example, the background may be distorted and things may be appearing where they shouldn’t be or wouldn’t naturally
- Blurred edges/unusual outline of face – as some deepfakes involve placing a digital face onto a person’s face, this may not be perfect and there is a chance you could spot it through the outline
- Check the source – think about where the content has arisen from. If it came from social media, it probably raises a few more red flags than via email or phone call initially
4 steps to reduce the risk of deepfakes
Take steps in your personal life
Protect your privacy by controlling who can access your online images. Consider making your social media profiles private to limit the accessibility of personal accounts, reducing the risk of misuse or fraud targeting you or your organization.
Educate and train your workers
It is important to make your employees more knowledgeable about the risks they face in your organization.
For example, with deepfake detector Sensity, the platform also offers interactive modules, real-world scenarios, and detailed training materials so workers can become well-equipped to spot manipulated media.
Encourage in-personal deals and conversations
If you suspect anything unusual in a conversation, prioritize in-person meetings to eliminate deepfake risks. For significant deals, a video call is more secure than a phone call, but meeting face-to-face remains the most reliable option to mitigate such threats.
Code word in business
Establishing a code word with suppliers, vendors, and third parties can help verify the authenticity of a request. This added layer of security ensures you can differentiate between legitimate communications and potential deepfake or fraud attempts.
Recap
Deepfake fraud attempts are increasingly common, posing significant reputational and financial risks. Trustpair offers robust protection against financial losses from cyber fraud. Equip your team to mitigate these threats by leveraging detection tools and enhancing employee training and awareness.
