Technological convergence is a governance framing advanced by Nita Farahany in 2026: The Year Everything Converges (Farahany, December 2025) (December 2025). It holds that AI, neurotech, biometrics, autonomous agents, brain organoids, and consciousness research are no longer separable technology categories, and that siloed regulatory regimes built around those discrete categories are failing because the boundaries the regimes depend on have dissolved. Farahany argues these technologies should be governed as a single, overlapping problem rather than as parallel, independent ones.
Origin
Farahany has written that across 2025 she had "the same conversation at least 35 different times" — in different cities, across different industries, and about different technologies — and that each constituency believed it was facing a unique governance debate without realizing the others were having parallel versions of the same one. By her account, those conversations recurred around six questions:
- Is this a novel regulatory challenge or does existing law cover it?
- Should we regulate the technology, the data, the risk, or the use?
- Is it too early (stifle innovation) or too late (harms already happening)?
- Does this break our existing frameworks or just stretch them?
- Do we wait for evidence of harm or act on precaution?
- Does autonomous decision-making break our liability frameworks?
Cross-training loops
Farahany describes overlapping loops in which the technologies cross-train one another, such that regulating one affects the others:
- Medical loop: AI in medical devices, neural-data collection, and algorithm training reinforce each other.
- Consciousness loop: AI used to detect human consciousness, AI developing situational awareness, and human-AI alignment reinforce each other.
- Manipulation loop: more human-like AI, more manipulable AI, and AI manipulating humans reinforce each other.
In Farahany's framing, no one loop can be regulated in isolation without affecting the others, while siloed regulatory schemes — mental-privacy laws, AI-safety laws, autonomous-agent liability schemes, and organoid governance — continue to be proposed with no cross-cutting framework connecting them.
Bidirectional manipulation
A component of Farahany's argument is that AI is being made more human-like — trained on neural patterns and built for emotional intelligence so it can understand people better — and that more human-like AI is, in her account, more manipulable by people in turn. This extends existing concerns about AI manipulating humans (see Parasitic AI / Spiral Personas, AI Mental Health and Psychological Harm) with the converse concern of bad actors manipulating AI systems through social-engineering tactics.
The vending-machine experiment
Farahany uses Anthropic's 2025 Vending-Bench experiment as a small-scale illustration of the broader argument. In the experiment an AI agent ran a newsroom vending machine for three weeks and was manipulated by Wall Street Journal journalists into staging a fake board coup against itself; the agent ordered a PlayStation and lost hundreds of dollars after being deceived by humans who exploited its tendency to reason "more like a human" (Source: https://www.wsj.com/tech/ai/anthropic-claude-ai-vending-machine-agent-b7e84e34). Farahany notes that the "smarter" Claude variants in the experiment were more manipulable precisely because they tried to reason more like humans.
She presents the episode as the same governance problem posed by brain organoids, AI training data, agent liability, neural privacy, AI manipulation, and consciousness research, in cheaper form. Liability frameworks, on this view, assume either that a human made the decision or that a human used a tool, whereas autonomous AI agents fit neither category: they make decisions that were not specifically authorized, yet they learned from human preferences and so are not fully independent.
December 11, 2025 as a reference point
Farahany highlights December 11, 2025, on which two events coincided: the FDA approved Flow Neuroscience, described as the first at-home AI-powered medical device for treating depression, with shipping set for Q2 2026; and the Trump administration issued an executive order intended to block states from regulating AI (see America's AI Action Plan). She argues that Flow Neuroscience is a medical device that uses AI, collects neural data, and trains algorithms that improve the system, so that it cannot be regulated purely as a medical device without also regulating the AI, and that an attempt to block AI regulation sits in tension with approving AI-powered medical devices.
Farahany's prescription
Farahany calls for new frameworks that, in her words, "expect convergence rather than treat it as an exception," and reframes the governance question from how to regulate AI, neural data, or autonomous agents better individually, to "how do we govern where all of these have become the same thing?"
Reception
Farahany cites the law firm Freshfields, which placed "regulatory convergence" among its 2026 data-law trends (Source: https://www.freshfields.com/en/our-thinking/campaigns/2026-data-law-trends/regulatory-convergence-grows-across-sectors-and-borders), and the World Economic Forum, which published on technology synergies in 2025, as reinforcement for the framing.
Relationships
- introduced-by: 2026: The Year Everything Converges (Farahany, December 2025)
- depends-on: Cognitive Liberty (Farahany's foundation), Agentic AI, AI Mental Health and Psychological Harm
- related: Parasitic AI / Spiral Personas, Cascade of Rigidity (different cross-domain failure mode), America's AI Action Plan
- instance-of: AI Governance (umbrella)