[SEAN-CSO 2026 HYBRID WORKSHOP] PeaceTech AI Lab: Turning P/CVE Challenges into AI Project Proposals

Artificial intelligence is opening new possibilities for P/CVE practitioners, from analysing online narratives to supporting decisions in day-to-day casework. Yet knowing what a tool can do is only the first step. Practitioners also need space to develop and test ideas against real problems, hear honest feedback, and shape a solution that fits their own context.

The SEAN-CSO 2026 webinar series explored how AI can support counter-narrative development, the detection of extremist narratives, practitioners’ decisions, and learning design. The PeaceTech AI Lab Hybrid Workshop builds on those conversations by moving from ideas to practice. Over four days, participants bring a problem from their own work and leave with a project proposal that integrates AI into a P/CVE initiative.

What the Workshop Aims to Achieve

The workshop is organised around two objectives:

1. Explore how AI and other digital approaches can contribute to addressing identified P/CVE challenges.
2. Develop an initiative or project idea that integrate responsible, feasible, and sustainable AI-based solution concepts.

The main output is a project proposal for integrating AI into a P/CVE initiative.

How the Workshop Works

Fifty participants join in a hybrid format, with 25 attending offline and 25 joining online. Each participant attends as a representative of their organisation, but develops an initial proposal individually.

Before the workshop, every participant completes a form describing their initial problem statement and proposed solution. The problem must relate to one of four problem areas set by the organisers. Participants are then grouped by how similar or relevant their problem statements and proposed solutions are, forming working groups of 8–9. Each group is supported by one AI expert.

Four Problem Areas

Every proposal starts from one of four problem areas. Each pairs a real context with a challenge framed as a “How might we…” question.

1. AI-Powered Early Warning System for Online Extremism

Southeast Asia’s highly connected and relatively young online population faces greater exposure to terrorist and violent extremist content. AI has potential to support the detection and analysis of online extremist activity, but its use also brings technical, political, and human-rights challenges. The challenge: how might we use AI to detect and understand emerging extremist narratives earlier, while accounting for the region’s diverse languages, cultures, and social contexts?

2. AI for Contextual Counter-Narratives

Responses to violent extremism need to account for local linguistic and cultural contexts. UNICRI’s regional work highlights the importance of adapting online P/CVE approaches to specific contexts and incorporating local knowledge and participation. The challenge: how might we use AI to help CSOs develop locally relevant counter-narratives that strengthen resilience against violent extremism across diverse Southeast Asian communities?

3. AI for Culturally Sensitive Online Safety

AI-based hate-speech detection remains challenging in Southeast Asian languages. A recent study covering Indonesian, Thai, Tagalog, and Vietnamese found that current multilingual models still struggle with low-resource languages, slang, implicit hate, and culturally nuanced expressions. The challenge: how might we develop AI solutions that can identify harmful or extremist content more accurately across Southeast Asian languages and cultural contexts, while reducing bias and misclassification?

4. AI-Assisted Support for P/CVE Practitioners

AI can support P/CVE through social-media monitoring, information analysis, and online-extremism detection. However, research also identifies challenges including algorithmic bias, public trust, and governance, which means AI should support rather than replace human decision-making. The challenge: how might we design an AI-assisted tool that helps P/CVE practitioners analyse digital information and make better-informed decisions, while keeping human judgement and community context at the centre?

Four Days: Feel, Do, Scrutinize, Pitch

The programme follows a clear flow. Participants first understand the problem, then build a solution, refine it with peers, and finally pitch it.

Day 1: Launch, Bonding, and Defining the Problem

The workshop opens with an opening ceremony and programme orientation. A family group session follows, using bonding activities to build trust and create a safe space where every member feels comfortable sharing.

In the afternoon, the Feel session, “What Bothers You?”, invites participants to share problems with their peer group, receive feedback, and develop clearer problem statements. Through writing, drawing, and reflection on key findings from the webinar series, participants identify concrete problems, needs, gaps, and opportunities for AI-based solutions. The day closes with Do: AI and P/CVE, where Irfan Amalee of PeaceGeneration Indonesia introduces the framework that will guide participants throughout the workshop.

Day 2: Learning from Experts and Creating the Solution

Day 2 begins with experts exploring how AI can be integrated into P/CVE work. Speakers include Matteo Vergani (Deakin University), Indrayanto Banyumurti (ICT Watch), and a representative of BNPT.

Participants then spend the main part of the day developing their solutions. Each participant works individually on a solution based on their validated problem statement and peer feedback, progressively refining the idea within their group. Six technical mentors provide feedback to help strengthen the proposed solutions. The design for change and system change frameworks guide this process.

Day 3: Scrutinise and Pitch

Day 3 focuses on scrutinising ideas and communicating their potential impact. Participants present their ideas, AI applications, expected impact, and prototypes to their peer groups, while practising active listening and giving constructive feedback. They also display their proposals on plano paper, allowing other participants to contribute written suggestions on sticky notes. Each group then selects one participant to pitch in the main room.

During the Pitching and Demo session, selected participants present their solutions and respond to questions from a panel and the audience. The panel includes Matteo Vergani (Deakin University), Priyank Mathur (Mythoslab), Asrul Daniel (SEARCCT), Raneeta Mutiara (SUSS), and SIGLA Research (Philippines).

Day 4: Awarding, Network Learning Space Launch, and Next Steps

The final day includes an awarding session recognising the participant or team with the most promising project, the official launch of the SEAN-CSO website, and a visit to Institut Teknologi Bandung (ITB).

The website launch marks an important step in activating a shared learning space for the SEAN-CSO network. Beyond serving as a knowledge repository, the platform is intended to help members access resources, revisit insights from the webinar series and workshop, and continue exchanging knowledge beyond the programme. It supports ongoing collaboration and collective learning, helping sustain the connections and ideas developed throughout the workshop.

Why Human Judgement Stays at the Centre

Across the four problem areas, one idea recurs: AI should support rather than replace human decision-making. The workshop’s focus on responsible, feasible, and sustainable solutions reflects this. Participants test their ideas with peers and mentors, and each problem area asks them to account for the languages, cultures, and communities involved, rather than treating technology as the answer on its own.

What Participants Will Take Away

• A project proposal that integrates AI into a P/CVE initiative, refined through peer and mentor feedback.
• Hands-on experience with the frameworks used in the workshop, including design for change and system change.
• Exposure to experts from academia, civil society, and government, as well as a panel that responds to their pitch.

Workshop Details

Peace AI Lab Hybrid Workshop 2026

• Format: Hybrid (25 offline and 25 online participants)
• Duration: Four days
• Date: 3–6 November 2026
• Venue: Deakin University Lancaster University Indonesia (Jl. Moch. Toha No. 77, Cigereleng, Regol, Bandung, West Java 40253)
• Online platform: Zoom Meeting

Follow the PeaceTech AI Lab Journey

See you in Bandung on 3–6 November! Until then, keep an eye on our Instagram and our website, seancso.org, for the latest updates on the PeaceTech AI Lab Hybrid Workshop.