What do you do when your foundations of reality are challenged? What do you do when you can’t spot fact from fiction? What do you do when you don’t know what’s real and what’s not? The rapid pace of development in artificial intelligence (AI) has shaken our perception of reality. We have gone from botchy Photoshop jobs to AI-generated images that look more authentic than photographs! This ability of AI has been leveraged by nefarious actors to make deepfakes.
Deepfakes are a new and increasingly sophisticated form of media manipulation used to create realistic and convincing images, audio, video, or text that appears to be real but isn’t. While deepfakes have primarily been associated with their impact on politics and entertainment, there are increasing risks to corporate businesses and reputations.
How Do Deepfakes Work?
Modern deepfakes are created using artificial neural networks (ANNs). ANNs are trained on a library of data, which allows them to extract patterns and trends to create the desired output. This training allows the ANN to learn how the subject looks or sounds under various conditions. For example, the ANN may learn how a person uses their expressions, intonation, and accent. Once it is trained, it can be used to create a deepfake. This is done by feeding the ANN new data that includes the desired subject’s face or voice. The ANN then uses the patterns and trends it learned from the training data to create a new output that looks or sounds like the desired subject.
Users can now create anything from abstractions of how the world will look in a hundred years to politicians in compromising situations.
Deepfakes used to be limited to convincingly altering or replacing the original content to create something spurious. However, with the advent of tools such as Midjourney, things have become even more serious. Users can now create anything from abstractions of how the world will look in a hundred years to politicians in compromising situations. These creations are often difficult to detect with the naked eye, making them a potent tool for spreading misinformation, damaging reputations, and causing widespread confusion.
Deepfakes in a Corporate Context
In a post-truth world, deepfakes are not a matter of levity. Increasingly, they are being used to extort money, manipulate the markets, and more.
-
Phishing and Fraud: Deepfakes can be employed in targeted phishing campaigns, where criminals impersonate high-ranking executives to trick employees into sharing confidential data or initiating financial transactions. The authenticity of the deepfake increases the chances of success, amplifying the financial and reputational risks. This has been a problem since 2019 when in one of the first corporate deepfake cases, a fraudster used AI voice technology to pose as the CEO of a German conglomerate. He then called the CEO of one of the firm’s subsidiaries and made him transfer $243,000 to a Hungarian bank account.
-
Insider Threats: Deepfakes can be used by disgruntled employees or insiders with malicious intent to create false evidence against the company, implicate innocent individuals, or extort sensitive information. Such attacks can lead to legal ramifications, damage intellectual property, and compromise trade secrets.
-
Brand Reputation: A well-executed deepfake can damage a company’s reputation overnight. False videos or audio clips featuring company executives endorsing unethical practices, spreading false information, or engaging in inappropriate behavior can tarnish the brand image and erode consumer trust. While criminals are yet to unleash the full force of deepfakes against corporates and businesses, individual brands and celebrities are already being targeted. For instance, AI tools have been used to imitate the voices of Leonardo DiCaprio, Quentin Tarantino, and George Lucas, among others, to make racist remarks.
-
Investor Confidence: Deepfakes have the potential to impact investor confidence, triggering volatility in stock prices and affecting the overall valuation of a company. Similarly, deepfakes of prominent investors may be used to lure others into investing in suspect businesses or scams. In a recent case, an AI-generated video of British finance journalist and investor Martin Lewis appeared to be calling for the public to invest in a new venture championed by Elon Musk, which turned out to be fake.
Tackling the Deepfake Dilemma
As the threat of deepfakes continues to evolve, corporates must adopt a proactive approach to safeguard their security and reputation.
-
Awareness and Training: Educating employees about deepfakes, their risks, and detection techniques is paramount. Regular training programs can help employees recognize and report suspicious content, reducing the likelihood of falling victim to deepfake-related attacks.
-
Advanced Detection Tools: Investing in cutting-edge deepfake detection technology is a necessary step. AI-powered algorithms and machine learning models can analyze video and audio files, identifying anomalies that indicate the presence of deepfakes. Collaborating with technology partners specializing in deepfake detection can enhance a company’s security infrastructure.
-
Establishing Verification Protocols: Implementing robust verification protocols for sensitive transactions, such as financial transfers or access to confidential information, can help mitigate the risks associated with deepfake-based phishing attacks. Multi-factor authentication and secure communication channels should be prioritized.
-
Crisis Response Plan: A comprehensive crisis response plan specific to deepfake incidents should outline steps to address and contain the situation, engage with stakeholders, communicate accurate information promptly, and restore trust in the company’s integrity.
-
Collaborating with Industry and Government: Corporations should actively engage with industry peers, technology providers, and government agencies to share best practices, exchange information on emerging threats, and collectively address the challenges posed by deepfakes. Collaborative efforts can lead to the development of standardized frameworks and regulations to combat this novel issue.
Avoiding Deep(er) Troubles
In a world where reality can be easily manipulated and truth becomes elusive, the rise of deepfakes presents a formidable challenge for corporate security and reputation. These manipulated videos, audio clips, and images have the potential to wreak havoc on businesses and their standing in the public eye.
Safeguarding security and reputation in the face of this new challenge requires a combination of vigilance, investment in advanced technologies, and collaboration. By staying informed, proactive, and responsive, businesses can navigate the complex landscape of deepfakes and preserve their integrity in the digital age. The battle against deepfakes is ongoing, but with the right strategies in place, corporations can effectively unmask these threats and protect their security and reputation.
____________
Written By: Nimesh Bansal
