Beeline AI People Search finds more people than volunteer reviewers

The LizaAlert chairman says that in 2024 a neural network found more people in drone images than the squad's group of more than 200 volunteer reviewers. He gives no counts.

  • 2 for an organization

    The chairman says that in 2024 the network found more people than the squad's group of more than 200 volunteer reviewers, a gain in capability. He gives no counts.

  • Not rated a person

    The chairman says that the technology helped save a person in 2021 but gives no details, so the depth for people found cannot be established.

Depth runs from 1, a convenience or time saved, to 5, a life or a livelihood preserved.

Status
Reported benefit.
Evidence

foundstudiedreportedclaimed

Only the makers and the deployer support the benefit.

Size

No magnitude yet: distribution and frequency are not documented.

Compared with

The squad's image-review group of more than 200 volunteers. Measured.

Without the system
Until 2024 the neural network found fewer people than the image-review group
With it
In 2024 the neural network found more people than the image-review group
The system’s part
Not isolated. The chairman credits the network with detections since 2019, a rescue in 2021 and the 2024 milestone. He says that it was trained on a large collected data set. Beeline's release says that it was retrained on 6,000 images in 2020. He gives no counts, so the size of the network's lead over the volunteer reviewers is unknown. No outside party has checked the comparison.
Not yet checked
LimitsCosts and harmsLearns from dataLater outcome

Download data Cite Suggest a correction

Benefit type
Increased Scalability
System
Beeline AI - Поиск людей (Beeline AI People Search), a neural network that finds people in drone photographs
Made by
Beeline (VimpelCom), with LizaAlert volunteers
Used by
LizaAlert volunteer search-and-rescue squad
Where
Russia
Stage
In routine use
Date
Kind
Aggregate result
Broad benefit
Physical health, Operational improvement
Last revised
1 November 2026

Sources

  1. News report, relays the account of the organization that used it.

    E1.RU, '«Люди перестали успевать»: как ИИ помогает «ЛизаАлерт» находить пропавших — нейросеть обогнала 200 добровольцев', 5 June 2026. The outlet relays the LizaAlert chairman's talk at the St. Petersburg International Economic Forum and contains no reporting of its own.

    В 2024 году произошло событие, которое для нас стало настоящим рубежом: нейросеть впервые нашла больше людей, чем группа просмотра снимков.
    Для ее обучения мы собрали огромный массив данных.

  2. Press release, by the makers.

    ComNews, 'В России реализован первый кейс по спасению потерявшегося человека с помощью решения "Beeline AI-Поиск людей"', 6 October 2020. Company news from Beeline, the system's developer.

    Поисково-спасательный отряд "ЛизаАлерт" нашел заблудившегося в лесу человека с помощью беспилотного летательного аппарата и нейросети "Beeline AI - Поиск людей" (0+), разработанной специалистами Билайн.

Details

AI system

Workflow

Squad volunteers flew drones that photographed large areas, and one photograph can show tens of thousands of square meters. A neural network trained on a large collected data set flagged the few pixels that a person occupies. Beeline joined the development and built the system into real search operations.

Follow-up

Conditions for replication

The squad has drone photographs in a volume that people cannot review in time. Volunteers act on the network's flags by sending a search group to the place.

Entry details

Sector

Search and rescue.

Purpose

Beeline built Beeline AI - Поиск людей for the LizaAlert squad to find people in drone photographs and to cut the time that volunteers spend reviewing them.

31 more questions not yet checked

model, release, provider, role, configuration, inputs and outputs, autonomy, human oversight, AI contribution, beneficiaries, reach, distribution and access, duration, every measurement question, every money question, ownership and control, dependencies, cross-border connections, geopolitical implications, replication attempts, availability, institutional decision, grounds, action and result, success criterion, every evidence question, observed or intended, alternative assessments, related entries, other identifiers, category mappings.

Cite this benefit

Peter H. Vartanian. (2026). Beeline AI People Search finds more people than volunteer reviewers (AIBD-1210). Artificial Intelligence Benefits Database, edition of 2 November 2026. https://petervartanian.xyz/aibd/records/AIBD-1210.html (archived at https://doi.org/10.7910/DVN/AADSII, version 9.0)

BibTeX Data Field definitions