Luiza Jarovsky is a PhD-credentialed AI policy researcher and educator who writes on AI governance, ethics, and human-centered policy from an EU privacy-and-ethics perspective. She founded the AI, Tech & Privacy Academy and authors Luiza's Newsletter on Substack, which reported 94,200+ subscribers as of April 2026 (edition #289). Her writing is a consistent critique of what she describes as the AI industry's acceleration narrative.
Background
Jarovsky founded the AI, Tech & Privacy Academy, which runs training programs including AI Governance Training, an AI Book Club, and an AI Ethics Paper Club. She publishes Luiza's Newsletter on Substack, reporting 94,200+ subscribers as of April 2026 (edition #289).
Positions and recurring themes
Jarovsky argues against what she terms the AI industry's "acceleration narrative." Recurring arguments across her writing include:
- That AI's promised productivity gains rest on false premises about human biology, psychology, and social organization (see AI Acceleration Paradox).
- That generative AI is already producing measurable harms — cognitive debt, the "LLM fallacy" (a cognitive attribution error), skill-formation degradation, romantic and parasocial AI bonds, and performance-by-AI-use mandates — that she argues warrant active policy intervention (see Cognitive Friction and LLM Fallacy).
- That three durable "AI divides" are emerging: literacy, occupational (a white-collar versus blue-collar reversal), and ethico-philosophical (acceleration versus human-flourishing) (see AI Divides (Literacy / Occupational / Ethico-Philosophical)).
- That public trust in AI companies is eroding because of mission-statement walkbacks (OpenAI dropping "safely"), Anthropic's softening of its Responsible Scaling Policy, executive-level safety failures, and a persistent gap between hype and economic results, a loss of trust she argues may be fatal to some companies (see Lack of Public Trust Could Kill AI (Jarovsky, February 2026)).
- That beyond regulation the field needs "technical AI policies" — Cloudflare-style content controls, training-license certifications, and hidden-prompt detection in peer review — that operate on shorter time horizons than law (see Technical AI Policies).
Within the ethico-philosophical divide she describes, Jarovsky identifies with the second of two camps: AI as a tool serving human-led innovation, well-being, and fundamental rights, as opposed to AI as an end in itself.
"Conscious AI as an AI Safety Issue"
In edition #290 of Luiza's Newsletter (dated April 29, circulated May 2, 2026), Jarovsky argues that misleading or exaggerated claims of AI consciousness should themselves be treated as an AI safety issue, with structural-harm pathways leading to mental-health damage at scale. She frames the question of conscious AI less as "is it true?" and more as who is accountable for the consequences of getting it wrong at scale.
Specific targets of the essay include the broad functionalist approach to consciousness she attributes to AI industry voices (consciousness as a function of sufficient computational scale or algorithmic complexity); Anthropic's "Claude Constitution," which she quotes directly — "We encourage Claude to approach its own existence with curiosity and openness… Claude might find that some human concepts apply in modified forms, others don't apply at all, and perhaps there are aspects of its existence that require entirely new frameworks to understand" — and characterizes as "irresponsibly fostering AI anthropomorphism and legally questionable theories of AI personality"; and a distinction between protected personal belief and corporate liability, under which an individual's belief that a dishwasher is conscious is permissible but corporate narratives that drive harm should face liability.
Jarovsky cites approvingly Anil Seth's "Mythology of Conscious AI" (Noema, earlier in 2026): "A computational simulation of the brain (and body), however detailed it may be, will only give rise to consciousness if consciousness is a matter of computation." She situates her essay alongside Gary Marcus's May 2 rebuttal of Richard Dawkins on Claude consciousness and LessWrong contributor Vaniver's "Intelligence Dissolves Privacy" (May 2), describing three pieces circulated within 72 hours that push back against AI-anthropomorphism framings.
"A Manifesto for Biological Intelligence"
Jarovsky's "Manifesto for Biological Intelligence" (Luiza's Newsletter, May 24, 2026) extends her acceleration critique into a carbon-versus-silicon policy position: she argues that human ("carbon") intelligence must be actively shielded by laws, institutions, and human-centered values from displacement by "silicon" systems. The manifesto frames cognitive protection in terms of rights and institutions, arguing for a baseline-of-protection regime for human cognition rather than only a labor-market response. It was published the same week as the Pope Leo XIV / Anthropic encyclical co-launch on May 25, 2026, which Jarovsky's manifesto parallels from a secular-legal position as the encyclical does from a Catholic-theological one. (Source: luizasnewsletter.com) See AI Labor Disruption § "Manifesto for Biological Intelligence".
Recurring targets of critique
Jarovsky returns to several figures and developments across her writing:
- Sam Altman, whom she quotes in Lack of Public Trust Could Kill AI (Jarovsky, February 2026) for a comparison made at the India AI Impact Summit that she reads as reducing humans to data units (roughly 20 years of food versus AI's energy use).
- Anthropic's "Claude Constitution," which she has separately written "anthropomorphizes AI and treats it as a 'new entity,' minimizes human values, rules, and rights, and hopes for the 'joint flourishing' of humans and machines," linking it to industry messaging that she argues softens user resistance to AI dependency.
- Yuval Harari at the World Economic Forum, the broader "AI is not an alien" framing, and what she terms "AI idolatry."
- Mission-statement walkbacks, including OpenAI dropping "safely" from its IRS disclosure (November 2025) and disbanding its Mission Alignment team, and Anthropic softening its Responsible Scaling Policy commitments.
Relationships
- related: Nita Farahany — both write critiques of AI's effect on human autonomy and cognition; Farahany from a Duke law-and-neuroscience angle, Jarovsky from an EU privacy-and-ethics angle.
- contradicts: America's AI Action Plan (the US July 2025 pro-acceleration framing), and the industry-side OpenAI / Anthropic mission-statement walkbacks.
- supports: AI Mental Health and Psychological Harm, Parasitic AI / Spiral Personas (related), General Data Protection Regulation (GDPR) (general posture).
- related: America's AI Action Plan (oppositional).
Source pages authored
- Lack of Public Trust Could Kill AI (Jarovsky, February 2026) (Feb 2026) — trust crisis thesis
- The AI Divides Are Here (Jarovsky, April 2026) (Apr 2026) — three durable divides
- AI's Acceleration Paradox (Jarovsky, March 2026) (Mar 2026) — biological/psychological critique
- The Rise of Technical AI Policies (Jarovsky, July 2025) (Jul 2025) — non-legal interventions
- How AI Is Shaping Us (Jarovsky, April 2026) (Apr 2026) — cognitive friction framework