The quest for authenticity in a digital world

THE potential for deepfakes to sway public opinion and influence the outcome of India’s Lok Sabha is raising red flags throughout the cyber community. While Indians are deciding on which candidate best represents their views, deepfakes and generative technologies make it easy for manipulators to create and spread realistic videos of a candidate saying or doing something that never actually occurred.

Deepfake threat in politics

The use of deepfakes in politics is particularly alarming. Imagine a scenario where a political candidate appears to be giving a speech or making statements that have no basis in reality. These AI-generated impersonations, based on a person’s prior videos or audio bites, could create a fabricated reality that could easily sway public opinion. In an environment already riddled with misinformation, the addition of deepfakes takes the challenge to a whole new level.

For instance, the infamous case where Ukrainian President Volodymyr Zelenskyy appeared to concede defeat to Russia is a stark reminder of the power of deepfakes in influencing public sentiment. Though the deception was identified due to imperfect rendering, there is no way of knowing who believes it to be true even after being disproved, showcasing the potential for significant political disruption.

A danger in the digital workplace

Employees, often the weakest link in security, are especially vulnerable to deepfake attacks. Employees could easily be tricked into divulging sensitive information by a convincing deepfake of a trusted colleague or superior. The implications for organizational security are profound, highlighting the need for advanced, AIdriven security measures that could detect anomalies in user behavior and access patterns.

Double-edged sword of AI in cybersecurity

However, it’s important to recognize that AI, the very technology behind deepfakes, also holds immense capabilities to help hackers discover cybersecurity loopholes and breach business networks. While AI may help discover new vulnerabilities for threat actors, it also could be used to discover countermeasures, such as identifying patterns in data that would have otherwise gone unnoticed.

A system could then flag the potential deepfake content and remove it before it achieves its goal. This could help bridge the global skills gap in cybersecurity, enabling analysts to focus on strategic decision-making rather than sifting through endless data.

Data dilemma

The proliferation of deepfakes feeds into the broader issue of fake news and bots, adding one more aspect to the inability of people to recognize legitimate sources from manipulated ones. The result of a news story by well-crafted AI and repeated by a deepfake could lead to public distrust or even incite mass unrest.

But let’s not forget that on the digital battlefield, AI is a weapon wielded by both defenders and attackers. Deploying algorithms to confirm unmanipulated data or discover mitigation efforts based on patterns in data could open new use cases for secure AI growth.

Regulatory-guided solution

Combating deepfakes requires a multifaceted approach that India lacks with its IT Act.

Legal frameworks specifically targeting the malicious creation and distribution of deepfakes are essential, along with international cooperation to manage the transnational nature of digital media. In the realm of technology and AI, ethical guidelines must be established to regulate the development and use of deepfake technologies. Media authentication frameworks, public awareness campaigns, and media literacy initiatives will be crucial in empowering individuals to distinguish between real and synthetic content. This collective effort is key to maintaining the integrity of digital media and the broader democratic process.

A business-first solution

The global call for regulating generative AI, including deepfakes, is growing. However, it’s important to recognize that comprehensive regulations primarily govern those within an industry, not individuals who operate outside legal boundaries.

Companies must prioritize AI-driven cybersecurity solutions as part of a broader, company-wide approach that intertwines safety with quality across all aspects of their operations. From online behavior to development processes, a centralized, AI-ingested understanding of an organization’s baseline is crucial. Such technologies could identify breaches in real time, whether perpetrated by external threat actors or employees misled by deepfakes. This proactive stance is essential for maintaining integrity and security in a digital landscape increasingly complicated by AI technologies.

Diwakar Dayal is the managing director and country manager at SentinelOne, a US-based cybersecurity company that delivers the defenses businesses need to prevent, detect and undo cyberthreats.