AI Policy Aspects of the Karnataka Gig Workers (Social Security and Welfare) Ordinance, 2025
- Abhivardhan
- 3 hours ago
- 5 min read
The Governor of the State of Karnataka had promulgated an ordinance entitled the Karnataka Gig Workers (Social Security and Welfare) Ordinance, 2025. This ordinance by design addresses certain labour law and employment-related health and safety issues associated with the ever-growing demographic of gig workers in the State of Karnataka.
While the ordinance focuses on primordial labour law and policy issues since gig workers may not receive similar benefits as "employees", there are some AI policy aspects around the contents of the ordinance, which merits a reasonable analysis of this ordinance.
Since this ordinance seems to be in force (post-promulgation), this explainer addresses the core AI policy and industry aspects around this ordinance, which is perhaps the first ever AI policy measure done by any Indian Statement at a legislative level, for it being comprehensive, and unique.
Application of the Ordinance
The applicability clause creates a comprehensive regulatory net with limited safe harbour protections, establishing three distinct categories of covered entities under the Karnataka Platform Based Gig Workers (Social Security and Welfare) Ordinance, 2025.
Territorial and Service-Based Safe Harbour Limitations
Clause (i) - Aggregator/Platform Coverage
No Geographic Safe Harbour: Any aggregator or platform "operating or providing" services within Karnataka falls under the ordinance, regardless of where they are incorporated or headquartered
Service-Specific Threshold: Only platforms providing services listed in the Schedule are covered, creating a limited safe harbour for platforms operating exclusively in unlisted sectors
Dynamic Expansion Risk: The phrase "as amended by the Government from time to time" might create the propensity to eliminate future safe harbour protections, as the government can expand coverage through notifications
Schedule-Listed Services (No Safe Harbour)
Ride sharing services
Food and grocery delivery services
Logistics services
e-Market place (B2B/B2C)
Professional activity provider
Healthcare
Travel and hospitality
Content and media services
Definitional Safe Harbour Gaps
Clause (ii) - Platform Definition Under Section 2(g)
The definition creates broad liability exposure with minimal safe harbour:
Covers any electronic system organizing work "at a particular location in return for payment"
Includes both automated monitoring and decision-making systems and human decision-making arrangements
No minimum transaction threshold or worker count exemption provides safe harbour protection
Registration-Based Safe Harbour Framework
Clause (iii) - Gig Worker Coverage
Conditional Application: Only applies to gig workers "registered with the Board under section 10"
Creates a perverse safe harbour where platforms might discourage worker registration to avoid obligations
Section 10(2) Mandatory Registration: Workers must register within 30 days, eliminating worker-choice safe harbour
Specific Provisions and Provisos Creating Safe Harbours
Limited Interim Safe Harbour - Section 21(4) Proviso | General Understanding |
"Provided that, until the operationalization of the Payment and Welfare Fee Verification System (PWFVS) by the Board, the Government may exempt such aggregator or platform who duly self report and submit details of the payouts made to their Gig workers in each of the transaction every quarterly." | Safe Harbour Conditions:
|
Administrative Cost Safe Harbour - Section 19(3) | |
"The Board shall not spend more than five percent of the fund to defray any administrative costs of the Board or employees of the Board." | This protects fund resources but doesn't create compliance safe harbour for platforms. |
Section 24 Proviso - Reporting Frequency Safe Harbour | |
"Provided that, the Government shall based on the Ease of doing business requirement may make the provision for submission of returns half yearly or annually by a notification." | Creates potential compliance burden relief for smaller platforms through reduced reporting frequency. |
No Safe Harbour Protections Seem to be Available For:
Welfare Fee Obligations (Section 20): 1-5% fee on all transactions with no minimum threshold exemption
Algorithmic Transparency (Section 13): Mandatory disclosure of automated decision-making systems affecting "prices, earnings, customer feedback"
Termination Procedures (Section 14): 14-day notice requirements with no emergency exception beyond "bodily harm" cases
Registration Deadlines: 45-day registration requirement for platforms with no extension provisions
Enforcement Safe Harbour Limitations
Section 23 - Penalties
No First-Time Violator Protection: Immediate ₹5,000 fine for first contraventions
Escalating Penalties: Up to ₹1 lakh for subsequent violations
12% Annual Interest: On delayed welfare fee payments with no grace period
The ordinance intentionally minimizes safe harbour protections, creating a strict liability framework that prioritizes gig worker protection over platform compliance flexibility. The only meaningful safe harbours are temporary (PWFVS exemption) or procedural (reporting frequency), rather than substantive liability protections.
Reference to Artificial Intelligence-enabled Automation Technologies
The best definition-level reference to AI is found in Section 2 of Ordinance, as provided in this excerpt below:
Figure 2 explains the 2 kinds of AI-based references in a mind map, with examples.
Section 7: Platform Based Gig Workers' Rights
AI Context: While Section 7 doesn't explicitly mention AI, it establishes fundamental rights that directly interact with "automated monitoring and decision-making systems" as defined in Section 2(a).
Safe Harbour Analysis
Rights-Based Safe Harbour Limitations
No Platform Discretion Safe Harbour: Section 7(a) mandates universal registration "irrespective of duration of work on any platform" - this eliminates any algorithmic filtering safe harbour that platforms might claim for worker eligibility
AI-Mediated Benefits Access: Section 7(b) creates conditional safe harbour based on "minimum number of transactions or gig work undertaken...in a quarter", but this threshold is determined by automated systems, creating potential AI bias vulnerabilities
Critical Gap: Section 7 provides no safe harbour protection against AI-driven rights violations. Workers have rights to registration and benefits, but no explicit protection when automated systems deny or manipulate these rights.
Section 13: Transparency in Automated Systems - The "AI Transparency" Framework
AI Definition Application: This section directly regulates "automated monitoring and decision-making systems" affecting "prices, earnings, customer feedback".
Safe Harbour Destruction Mechanisms
Algorithmic Disclosure Obligations
No Trade Secret Safe Harbour: Section 13(1) requires disclosure of "automated monitoring and decision-making criteria" with no proprietary algorithm exemption
Language Accessibility Mandate: Information must be provided "in simple language and in Kannada, English or any other language listed in the Eighth Schedule".
Section 13(1) establishes a two-tier transparency obligation that fundamentally restructures the relationship between platforms and gig workers regarding AI decision-making systems.
Tier 1: Procedural Transparency
Platforms must inform workers about "the procedure to seek information" regarding automated parameters
This creates a mandatory disclosure pathway rather than voluntary platform discretion
Workers gain procedural rights to access algorithmic information, not just outcomes
Tier 2: Substantive Parameter Disclosure
Covers "automated monitoring and decision making parameters" affecting working conditions
Explicitly includes "fares, earnings, customer feedback and allied information"
Extends to any parameters that "have an impact on their working conditions"
Strategic Logic Behind This Approach
Accessibility Over Technical Complexity
The requirement for "simple language" in local languages (Kannada, English, or Eighth Schedule languages) ensures:
Workers can actually understand algorithmic explanations
No hiding behind technical jargon or English-only documentation
Democratization of algorithmic knowledge
Procedural Rights Create Ongoing Accountability
Rather than one-time disclosure, the section creates an ongoing right to seek information, meaning:
Workers can request explanations when algorithms change
Platforms cannot claim "trade secrets" to avoid transparency
Creates continuous monitoring pressure on algorithmic fairness
Comprehensive Coverage of AI Impact Areas
The inclusion of "fares, earnings, customer feedback" targets the most economically sensitive algorithmic decisions:
Fares: How pricing algorithms affect worker income potential
Earnings: How payment calculation systems operate
Customer feedback: How rating/review systems impact worker opportunities
Allied information: Catch-all for related algorithmic impacts
Anti-Discrimination Requirements
Absolute Liability Standard: Section 13(2) requires platforms to "take measures to prevent discrimination through automated monitoring and decision-making systems based on religion, race, caste, gender, or disability"
No Intent Safe Harbour: The provision creates strict liability - platforms cannot claim algorithmic discrimination was unintentional.
Critical Enforcement Gap: While Section 13 eliminates transparency safe harbours, it provides no specific enforcement mechanism or penalties for non-disclosure of algorithmic parameters.
Anyways, these are the most important and directly impacting AI policy aspects of the Karnataka Government's Ordinance promulgated recently. In case you have any confusion, please email us at vligta@indicpacific.com to discuss further.