RFP/PBF/2302/03 - Use of AI-augmented tools and machine learning

UNDP
RFP/PBF/2302/03 - Use of AI-augmented tools and machine learning Request for proposal

Reference: UNDP-LKA-00052
Beneficiary countries or territories: Sri Lanka
Registration level: Basic
Published on: 23-Feb-2023
Deadline on: 01-Mar-2023 03:30 (GMT -5.00)

Description
Use of AI-augmented tools and machine learning to improve approaches to monitoring and analysing digital harms on social media in Sri Lanka. 

Both through the ongoing project supported by the UN Peacebuilding Fund and prior peacebuilding interventions such as the Joint Programme for Peace (JPP) and Preventing Violent Extremism through Promoting Tolerance and Respect for Diversity, the United Nations Development Program (UNDP) in Sri Lanka has made headway in understanding hate speech patterns online. At present, UNDP Sri Lanka is supporting human-centric monitoring efforts on social media platforms including Facebook, YouTube and TikTok. However, this approach is not without its limitations. Recognizing this, UNDP Sri Lanka believes that increased investment is needed to augment these efforts with automated processes. Specifically, it wishes to work with an organization to explore ways to leverage machine learning, natural language processing, data mining and artificial intelligence techniques to identify, collate, categorize and analyze publicly available dangerous content on pre-selected social media platforms that are popular in Sri Lanka (at least covering Facebook, TikTok and Youtube) in both Sinhala and Tamil. 

OBJECTIVE AND SCOPE OF INTERVENTION 

The purpose of the initiative is to utilize machine learning, natural language processing, data mining and artificial intelligence techniques to systematically identify, collate, categorize and analyze publicly shared hate speech on selected social media platforms in Sri Lanka. The results are expected to provide an understanding of the prevalence of hate speech, key drivers, and narratives. At minimum, the techniques in question should effectively navigate information from public posts shared on Facebook in Sinhala and Tamil. Preferably, the techniques should also be able to effectively parse trilingual data from video-centric platforms too—with YouTube and TikTok being of particular interest in the Sri Lankan context. The initiative is not expected to draw insights from Twitter, without explicit UNDP prior approval, as the local userbase on the platform remains small.

The initiative will include the development of monthly datasets and analytical briefs, based on the information gathered through the aforementioned techniques.

Interested suppliers must submit their offer directly in the system as instructed in the solicitation document, following the instructions in the available user guide. if you have an account, log in in this link: http://supplier.nextgenerp.partneragencies.org/ using your username and password. User the forgotten password/username feature if you do not remember them. If you do not have an account already, you can register one following this link: https://estm.fa.em2.oraclecloud.com/fscmUI/faces/PrcPosRegisterSupplier?prcBuId=300000127715271.

Search for the specific tender using search filters and subscribe to the tender in order to get notifications in case of amendments of the tender document. If you need support with the online system, you can contact the contact details of this tender as indicated in the solicitation document.

UNDP Sri Lanka
Procurement Unit

Email address: procurement.lk@undp.org
Surname: Procurement Unit