AI is a double-edged weapon, one that empowers both perpetrators and protectors.
Vaishnaw warns against misinformation, disinformation and deepfakes
Robust “techno-legal framework” can balance AI's opportunities with its risks
“Privacy cannot be compromised.”
These resolute words from Senior Advocate Vivek Sood at the India AI Impact Summit 2026 encapsulated the profound tension at the heart of the gathering's discussions on artificial intelligence and security. Delivered during the session “AI for Secure India: Combating AI-Enabled Cybercrime, Deepfakes, Darkweb Threats & Data Breaches” on February 17, 2026, at Bharat Mandapam in New Delhi, Sood's statement grounded the conversation in constitutional principles while acknowledging the urgent realities of evolving digital threats. With more than three decades at the Bar, predominantly representing the accused, Sood offered a defense counsel's perspective, reminding participants that the pursuit of technological solutions to cybercrime must never override the presumption of innocence.
The fundamental right to privacy enshrined under Article 21 of the Indian Constitution and affirmed in the landmark Puttaswamy judgment.
The India AI Impact Summit 2026, hosted under the IndiaAI Mission by the Ministry of Electronics and Information Technology from February 16 to 20, positioned itself as a defining moment for global AI discourse, the first such major event led by a Global South nation. Bringing together heads of state, leaders from companies like NVIDIA, Anthropic, and OpenAI, policymakers, researchers, and legal experts, the summit explored AI's potential across themes of People, Planet, and Progress while placing heavy emphasis on safe, trusted, and inclusive systems. The dedicated cybercrime session, supported as knowledge partner by the Future Crime Research Foundation, featured a panel including Sood, alongside experts such as Rakesh Maheshwari, a former senior official at MeitY instrumental in shaping India's cyber laws and data governance, and other voices from law enforcement, industry, and academia. Their collective insights painted AI as a profoundly ambivalent force—capable of fortifying defenses against fraud yet simultaneously amplifying criminal sophistication.
Sood began by addressing direct questions on whether AI could effectively “hunt” fraudsters and counteract deepfakes, dark web operations, and large-scale data breaches. He reframed the inquiry from the standpoint of someone who often stands for “the accused, the condemned man,” cautioning against an overly prosecutorial mindset.
Privacy, he insisted, stands as non-negotiable, a core right that flows from life and liberty itself. While crime prevention and investigation represent recognized exceptions, any use of AI in law enforcement demands careful calibration to prevent erosion of civil liberties. He described artificial intelligence as “a double-edged weapon,” one that empowers both perpetrators and protectors. Criminals exploit it for automated scams, hyper-realistic deepfakes that deceive victims or manipulate narratives, and anonymized dark web activities. Conversely, law enforcement can leverage AI for real-time pattern detection in financial flows or communications, enabling prevention rather than mere reaction.
Sood particularly highlighted the practical shortcomings of the existing criminal justice system in addressing low-value cyber frauds, where victims often find that pursuing recovery proves economically unviable despite their desperation for restitution and punishment. Preventive AI frameworks that flag suspicious activities before harm occurs could, he suggested, offer a more humane and efficient path forward. Victims seek two things above all, yet the digital realm severely complicates attribution. Unlike physical crimes where perpetrators leave traceable footprints, cyber offenses rely on layered service providers, VPNs, and cross-border anonymity that render pinpointing originators extraordinarily difficult. Sood invoked the enduring doctrine of criminal conspiracy in the Indian Penal Code, unchanged since 1860, which permits inference of coordinated intent from circumstantial actions without needing explicit proof of agreement. The law itself, he maintained, possesses the flexibility to adapt to new criminal innovations; the genuine obstacle lies in execution—jurisdictional barriers, sluggish international cooperation through mutual legal assistance treaties and Interpol, and the sheer scale of offenses that leave most individual victims beyond effective redress. In such a landscape, he noted, following the “money trail” through laundering networks remains among the most powerful investigative instruments available.
These concerns found resonance across the summit. Union Minister Ashwini Vaishnaw repeatedly underscored the deepening crisis of deepfakes and disinformation, stating that “we need much stronger regulation on deepfakes. It is a problem growing day by day,” and warning that “misinformation, disinformation, deepfakes, they are attacking the foundation of society.” He further declared that “innovation without trust is liability,” emphasizing the necessity of international collaboration to safeguard societal trust and institutional integrity against these systemic threats. India's Principal Scientific Adviser Ajay Sood advocated for a robust “techno-legal framework” to balance AI's opportunities with its risks, describing the technology as a double-edged sword that requires governance to maximize benefits while minimizing harms like misinformation. Global Cyber Alliance Interim CEO Brian Cute highlighted how generative AI has made cybercrime “cheaper and more effective,” driving exponential increases in phishing, scams, and attacks coordinated from centers in Southeast Asia.
Former CERT-In Director General Gulshan Rai stressed the critical importance of securing AI algorithms and underlying systems themselves, warning that unaddressed vulnerabilities could lead to cascading failures. Broader summit conversations addressed AI's implications for democracy, including its weaponization for hate speech, surveillance, and election interference, often disproportionately affecting minorities. Pre-summit workshops, such as those organized by the CyberPeace Foundation, trained professionals in detecting and countering misinformation, deepfakes, and fake news to bolster digital trust and ethical practices.
Prime Minister Narendra Modi framed AI as a “force for good,” essential for strengthening public systems, promoting sustainable development, and ensuring inclusive growth through safe deployment. The event featured expansive exhibitions with over 300 participants from more than 30 countries, thematic pavilions, and initiatives like funding for deepfake detection research at institutions such as the IITs, alongside a “DPI approach” to democratize AI resources and significant investments including a $1.1 billion state-backed venture fund for AI startups. Yet experts cautioned about challenges ranging from workforce displacement to bioterrorism risks amplified by AI, underscoring the need for stronger global partnerships.
In closing his remarks, Vivek Sood articulated guiding principles for a secure digital India: the establishment of a deterrent legal framework to curb cybercriminality at its source, the firm aspiration that no crime should escape punishment, and a vital caution against obsessive prosecution that risks turning the justice system into a one-way street devoid of protections for the innocent accused.
As the session concluded, it affirmed India's emerging leadership in AI governance, demonstrating that true progress demands anchoring technological advancement in constitutional morality, procedural fairness, and the rule of law. In an age where AI accelerates both extraordinary promise and unprecedented peril, the voices from this session—Sood's insistence on uncompromised privacy, Vaishnaw's call for trust as the bedrock of innovation, and the panel's shared recognition of AI's dual nature, served as a compelling reminder that security and liberty must advance together, never at each other's expense.
