evidoria

← Back to browse

Good practice Imported

Sama's AI 101 Training & Job Referral Programme — RCT Shows 37% Earnings Gains for Nairobi Youth

Kenya · Nairobi · See the Kenya profile · See the Nairobi profile

Evidence: Randomised controlled trial Top 56% 47/100 · Ask Evidence Copilot about this practice

A randomised trial of Sama's 'AI 101' data-annotation training combined with a guaranteed job referral raised monthly earnings by 37% and cut unemployment by 10 points among low-income Nairobi youth, sustained 2.5 years later — training alone had no effect.

+37%
Monthly earnings gain, training + referral (vs KES 13,440 baseline)
-10 percentage points
Unemployment reduction, training + referral
+60%
Earnings gain for women
+26%
Earnings gain for men
1,115
Applicants randomly assigned

Details

Maturity
Established
Promoter
Sama (SamaSource), evaluated by Innovations for Poverty Action / J-PAL
Period
2017-2020 RCT; effects tracked to 2.5 years post-training
Keywords
AI data annotation, workforce training, youth employment

Context

Sama (formerly Samasource) runs "AI 101," a short vocational course teaching low-income young Kenyans the basics of data annotation and image/text labelling used to train machine-learning systems, in Nairobi. Researchers from Innovations for Poverty Action randomly assigned 1,115 applicants aged 18-35 to training only, training plus a guaranteed referral to a paid data-annotation job at Sama, or a comparison group.

Results

Training combined with a job referral raised monthly earnings by 37% (from a KES 13,440 baseline) and cut unemployment by 10 percentage points; effects were driven almost entirely by participants who obtained a Sama placement, and were more than twice as large for women (60% earnings gains) as for men (26%).

Conclusions

Training alone, without the referral, produced no measurable earnings gains and was associated with lower earnings at 16 months, underlining that skills training without a guaranteed employment pathway may not pay off on its own. Sama's broader content-moderation business has faced separate, well-documented labour-practice criticism, including a 2022 Kenyan lawsuit over alleged low pay and inadequate psychological support for Meta content moderators.

Implementation
Indicative cost
Low (< €50k)
Time to results
Medium (1–3 years)
Staffing & skills
Sama (formerly Samasource), Innovations for Poverty Action researchers

Conditions for success

  • training paired with a guaranteed job referral to a paid placement is the key differentiator versus training alone

Common failure modes

  • training alone without a guaranteed referral produced no measurable earnings gains and was linked to lower earnings at 16 months; Sama's wider labour practices have faced separate criticism, including a 2022 Kenyan lawsuit

Commonly funded by

National / regional programmes ESF+ — European Social Fund Plus

Indicative funding routes for practices of this type — always check each programme's current calls and eligibility rules.

Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.

Data sources

Where this practice's information was retrieved from, and when.

Similar practices you may find useful