The Artificial Intelligence Bill, 2026

19 Feb 2026 domainSenate Senate Bills No. 4 of 2026 Gazette: Supplement No. 15
View original PDF

Status Timeline

Proposed

Active now

Public Review

Pending

Debate

Pending

Executive

Pending

Final

Pending

1

Establishment of the Artificial Intelligence Commissioner

Establishes the independent Office of the Artificial Intelligence Commissioner as a State Office under Article 260(q) of the Constitution (Section 4).

Key responsibilities include:

  • Overseeing the enforcement of the Act and conducting risk assessments.
  • Establishing and managing regulatory sandboxes to test AI systems under oversight.
  • Maintaining a public register of high-risk AI systems.
  • Investigating complaints concerning bias, discrimination, or rights violations.
2

Four-Tier Risk Classification System

Categorises AI systems into four risk levels under Section 25:

  • Unacceptable Risk: Systems that pose severe threats, which are strictly prohibited.
  • High Risk: Systems used in critical sectors including healthcare, education, agriculture, finance, security, employment, or public administration.
  • Limited Risk: Systems with moderate risks.
  • Minimal Risk: Systems with negligible risks.
3

Mandatory Assessments for High-Risk AI

Imposes strict obligations on providers and deployers of high-risk AI systems (Section 26):

  • Must conduct both risk and human rights impact assessments before deployment.
  • Must maintain records of data inputs, training datasets, and performance metrics for at least five years.
  • Must ensure transparency, traceability, and explainability of decision-making processes.
  • Must obtain explicit consent from individuals before generating or manipulating synthetic media (images, voice, or likeness) and label the output clearly as AI-generated.
4

Workforce Impact Obligations

Under Section 33, any provider or deployer of an AI system likely to impact employment must:

  • Conduct a workforce impact assessment to determine potential job displacement.
  • Implement mitigation measures, including reskilling programs, in collaboration with national and county governments.
5

Fines Up to KSh 5 Million

Establishes a dual-tier penalty structure under Section 35:

  • Major Offences: Deploying prohibited AI, failing to perform required high-risk assessments, failing to conduct workforce impact assessments, or generating non-consensual synthetic media carries a fine of up to KSh 5 million, imprisonment for up to two years, or both.
  • Other Offences: Failing to comply with transparency obligations, violating ethical guidelines, or obstructing the Commissioner carries a fine of up to KSh 1 million, imprisonment for up to six months, or both.

Comments (0)

No comments yet. Be the first to start the discussion!