In an interview, Google CEO Sundar Pichai argues against knee-jerk regulation and blanket vetting of AI algorithms, advocates for sector-by-sector approach
Tim Bradshaw / Financial Times :
Context & Ripple Effects
This 2019 Financial Times interview is the opening move in Sundar Pichai's multi-year public positioning on AI governance. At this point he warns against knee-jerk rules and blanket algorithm vetting, favoring a sector-by-sector approach.
The later record shows him closing the loop without abandoning the substance: by January 2020 he says AI is too important not to be regulated, by December 2020 he argues GDPR proves regulation can get it wrong, and by October 2021 he is explicitly calling for new US federal privacy legislation. Read together, the consistent thread is Google endorsing regulation on its own preferred terms — broad principles, targeted per-sector rules — rather than ex-ante review of models themselves.
First-order effects
- Regulators drafting early AI frameworks receive a concrete counter-proposal from the field's most prominent company: regulate applications sector by sector, not algorithms before deployment.
- Google's own AI-heavy products — Search, Assistant, Android — are the direct beneficiaries, since blanket vetting would subject them to pre-approval while sector-specific rules leave them governed only where they touch regulated domains.
Second-order effects
- Rival cloud and consumer-AI players face pressure to align with the sector-by-sector framing or argue for stricter alternatives, turning regulatory design into competitive positioning among hyperscalers.
- If lawmakers adopt application-level rules over algorithm-level vetting, compliance costs shift from model developers broadly onto specific regulated industries — healthcare, finance, transport — reshaping who bears the burden of AI oversight.
Third-order effects
- The pattern across these interviews suggests large AI labs will keep steering toward a regime of broad statutory principles plus industry input — the 'state-compatible' posture — rather than accepting mandatory pre-deployment review of frontier systems.
- Whether that holds depends on whether high-profile AI failures force governments toward the blanket-vetting model Google rejects; the sector-by-sector approach persists only as long as incidents stay contained within individual sectors.
The trend: Google's AI-policy line has evolved from cautioning against premature rules in 2019 to championing comprehensive but company-shaped regulation, with Pichai positioning the firm as co-author of whatever oversight emerges.