AI’s double bind in Africa and South Asia: repression tools, political deepfakes, and the “train allies, deny rivals” playbook

A new wave of reporting links artificial intelligence to coercive governance and political manipulation across regions. One article argues that AI is becoming a “new machinery” for African repression, implying faster surveillance, more scalable intimidation, and improved targeting of dissent. In parallel, another story centers on India’s Uttar Pradesh Congress chief Ajay Rai, who refuted allegations tied to an AI-generated video that purportedly contained a derogatory remark about the prime minister. The dispute highlights how synthetic media can rapidly become a political weapon, forcing parties to fight on both narrative and evidentiary ground. A third piece from RAND frames a strategic logic—“Train Thy Neighbor, Deny Thy Adversary”—suggesting that states can shape security outcomes by building partner capacity while constraining rival influence. Geopolitically, the cluster points to a convergence of information operations, internal security, and partner-force strategy. If AI meaningfully lowers the cost of monitoring and repression, governments gain leverage over opposition while raising the risk of rights backsliding and international reputational costs. The India deepfake controversy shows how synthetic content can destabilize domestic politics and complicate crisis communication, especially when social platforms amplify unverified clips. RAND’s partner-training logic implies that external actors may increasingly treat AI-enabled intelligence and influence as part of security cooperation, not just technology policy. Overall, the winners are actors that can move faster than scrutiny—security services, political machines, and aligned information ecosystems—while the losers are institutions that rely on slow verification, independent media, and trust-building. Market and economic implications are indirect but potentially material through security spending, platform risk, and insurance/defense procurement. AI-driven repression narratives can lift demand for surveillance, cyber monitoring, and “information integrity” tooling, supporting defense-tech and cybersecurity budgets in affected countries and regions. The political deepfake episode can increase compliance and moderation costs for social media and video platforms, and it can raise volatility in reputational-risk-sensitive sectors such as advertising and consumer trust-driven businesses. RAND-style partner training can also translate into procurement flows for training services, communications equipment, and intelligence support—affecting defense contractors and logistics providers. While no specific commodity or currency moves are stated in the articles, the direction is toward higher risk premia for digital trust and higher capex/opex for security and verification infrastructure. What to watch next is whether these narratives translate into concrete policy and procurement decisions. In Africa, look for evidence of expanded surveillance authorities, new AI procurement contracts, and tighter controls on protest coordination and online organizing. In India, monitor how courts, election commissions, and political parties establish standards for authenticating synthetic media, including whether forensic watermarking or platform takedown protocols become formalized. For the RAND “train/deny” approach, track announcements of security assistance packages, joint training curricula, and any explicit references to AI-enabled intelligence, targeting, or influence operations. Trigger points include high-visibility viral synthetic-media incidents, legislative moves on digital regulation, and partner-country capacity-building milestones that signal a shift from general cooperation to AI-augmented security posture.

Geopolitical Implications

• 01

  AI lowers the operational cost of surveillance and coercion, potentially strengthening authoritarian control while increasing international friction.

• 02

  Deepfakes can erode domestic political trust and complicate crisis communication, raising the risk of reactive governance and polarization.

• 03

  Security assistance strategies may shift toward capacity-building that explicitly targets information and influence domains, not only conventional defense.

Key Signals

• —New digital regulation or election-media authentication rules addressing synthetic content
• —Forensic watermarking, provenance standards, or platform takedown protocols becoming mandatory
• —Procurement announcements for surveillance, monitoring, and AI-enabled analytics tied to internal security
• —Security assistance packages referencing AI-enabled intelligence, targeting, or information operations