"Blaming the human condition lets technology and power off the hook." Discuss the ethical responsibility of AI developers and states for the social consequences of AI.

The statement rejects fatalism, the view that deepfakes, misinformation and synthetic content are simply part of being human. These harms come from design and deployment choices. The duty to answer for them therefore lies first with those who build AI and those who govern it.

Why fatalism fails ethically

• It moves the burden onto individuals and away from institutions. That breaks the principle that power entails accountability.
• AI makes it cheap to mass-produce believable content. Even peer-reviewed journals have published AI-written papers that slipped past plagiarism checks, which weakens trust in shared facts.

Responsibility of AI developers

• Safety by design: build in guardrails against foreseeable misuse such as impersonation and non-consensual intimate imagery [2].
• Provenance: support clear labelling and traceable metadata for synthetic content [2].
• Transparency and explainability: people affected by AI outputs should be able to understand how they were produced, as UNESCO's 2021 Recommendation requires [4].
• Authorship ethic: a responsible human should stand behind every output and be able to explain it.

Responsibility of the state

• Protective duty: Article 21 (dignity, privacy) requires the state to act against AI harms. Existing law already punishes obscene deepfakes under IT Act Sections 67A/67B [1].
• Platform accountability: the IT Rules, 2021 were amended on 10 February 2026 to cover synthetically generated information. They impose labelling and due-diligence duties on intermediaries [2].
• Proportionate governance: the India AI Governance Guidelines (2025) set out a risk-based framework of "innovation with safeguards" [3].
• Civic capacity: law can only remove unlawful content. Misleading but lawful content needs media and information literacy for citizens [4].

Limits and shared responsibility

• Over-regulation can chill free speech under Article 19(1)(a) and slow innovation. Restrictions must be necessary and proportionate.
• Users, journals and the media also carry duties: to verify before sharing and to strengthen their own checks.

Concern about AI is real, but treating it as fate lets the actors best placed to fix it do nothing. Developers who build safety into their systems, a state that regulates in proportion to risk, and citizens who can judge what they read together turn that concern into accountability. This is the human-centric AI that constitutional values and UNESCO's ethics framework envisage.