Media

Interviews, press coverage, and videos.

Reflecting on the Workshop
  LinkedIn Post

Reflecting on the Workshop

Recap of the two-day AI Interpretability & Safety Workshop at IISc Bengaluru, organized by Agyeya Research, bringing together 60+ students and researchers with practitioners from Google DeepMind, Microsoft, and IBM; Pulkit Verma presented "Beyond the Lab: User-Aligned Assessment of Taskable AI Systems."

Agyeya Research  · Sep 20, 2026
AI Futures Dialogue #1: Beyond the Lab
  LinkedIn Post

AI Futures Dialogue #1: Beyond the Lab

Announcement for the first AI Futures Dialogue session, part of the Technical AI Governance Reading Group, featuring Pulkit Verma presenting "Beyond the Lab: User-Aligned Assessment of Taskable AI Systems."

Secure AI Futures Lab  · Aug 19, 2026
Emerging AI Technologies for a Sustainable Future: Prospects and Challenges
  LinkedIn Post

Emerging AI Technologies for a Sustainable Future: Prospects and Challenges

Agenda announcement for a one-day hybrid workshop with WRI India on AI's role in sustainability, featuring Pulkit Verma as a panelist on the impact of AI on the Indian economy and environmental sustainability.

Department of HSS, IIT Tirupati  · Apr 10, 2026
AI Safety Connect Day: Losing Control of Advanced AI Systems?
  LinkedIn Post

AI Safety Connect Day: Losing Control of Advanced AI Systems?

Recap of an AI Safety Connect Day panel in New Delhi on whether we will lose control of advanced AI systems, featuring Pulkit Verma alongside Stuart Russell, Sarah M. Erfani, Kee-Eung Kim, and Adam Gleave.

AI Safety Connect (AISC)  · Mar 10, 2026
Teaching Robots to Handle Tomatoes (....and more!)
  YouTube Short

Teaching Robots to Handle Tomatoes (....and more!)

A short video from MIT CSAIL Alliances featuring robot manipulation research.

MIT CSAIL Alliances  · Jan 27, 2026
Hosting ISRO and Chanakya University at MIT AeroAstro
  LinkedIn Post

Hosting ISRO and Chanakya University at MIT AeroAstro

MIT AeroAstro hosted ISRO and Chanakya University's Dr. Somanath Sreedhara Panicker and Dr. Yashavantha Dongre for a discussion with faculty and students and lab tours, with Pulkit Verma among the attendees.

MIT AeroAstro  · Oct 15, 2025
How to teach AI models to plan?
  Press

How to teach AI models to plan?

MIT researchers demonstrate that large language models can improve planning accuracy by writing out each step explicitly and correcting errors flagged by automated verification systems.

PRESS Insider  · Sep 25, 2025
Interview with Pulkit Verma: Towards safe and reliable behavior of AI agents
  Press

Interview with Pulkit Verma: Towards safe and reliable behavior of AI agents

A doctoral researcher discusses his work on enabling third-party assessment of black-box AI systems and personalizing explanations of AI behavior to enhance user understanding and trust.

AIhub  · Oct 24, 2024
New AI symposium helmed by Siddharth Srivastava connects experts and industry
  Press

New AI symposium helmed by Siddharth Srivastava connects experts and industry

ASU's School of Computing and Augmented Intelligence co-organized a Stanford symposium focused on safety assessment of adaptive AI systems that learn and adapt to diverse users.

ASU SCAI News  · Apr 2, 2024
Robot coworkers: How AI impacts the future of work
  Press

Robot coworkers: How AI impacts the future of work

Researchers at ASU are developing AI-equipped autonomous robots with built-in tutoring systems to help factory workers adapt to automation.

ASU News  · Nov 7, 2019

Contributed Talks

AAAI 2025 Tutorial on User-Driven Capability Assessment of Taskable AI Systems
  YouTube Video

AAAI 2025 Tutorial on User-Driven Capability Assessment of Taskable AI Systems

Recorded tutorial on methods for enabling users to assess and understand the capabilities and limits of taskable AI systems.

AAAI 2025  · Mar 14, 2025
Epistemic Exploration for Generalizable Planning and Learning in Non-Stationary Settings
  YouTube Video

Epistemic Exploration for Generalizable Planning and Learning in Non-Stationary Settings

Talk on an interactive approach for discovering a black-box sequential decision-making agent's capabilities in stochastic, non-stationary settings.

ICAPS 2024  · Jun 13, 2024
Autonomous Capability Assessment of Black-Box Sequential Decision-Making Systems
  YouTube Video

Autonomous Capability Assessment of Black-Box Sequential Decision-Making Systems

Talk on an active-learning approach for modeling the capabilities of black-box AI systems that can plan and act in stochastic settings.

ICAPS 2023 Workshop on Knowledge Engineering for Planning and Scheduling (KEPS)  · Jul 8, 2023
Discovering User-Interpretable Capabilities of Black-Box Planning Agents
  YouTube Video

Discovering User-Interpretable Capabilities of Black-Box Planning Agents

Talk on an algorithm for discovering the high-level, user-interpretable capabilities of an AI system with arbitrary internal planning algorithms.

KR 2022  · Jul 31, 2022
Differential Assessment of Black-Box AI Agents
  YouTube Video

Differential Assessment of Black-Box AI Agents

Talk on an approach for differentially assessing black-box AI agents that have drifted from their previously known models.

AAAI 2022 Workshop on Artificial Intelligence Safety (SafeAI)  · Jun 20, 2022
JEDAI: A System for Skill-Aligned Explainable Robot Planning
  YouTube Video

JEDAI: A System for Skill-Aligned Explainable Robot Planning

Talk on JEDAI, a system that helps non-experts create high-level, intuitive robot plans while ensuring they are executable, with explanations of errors.

ICAPS 2022 Workshop on Explainable AI Planning (XAIP)  · Jun 13, 2022
JEDAI: A System for Skill-Aligned Explainable Robot Planning (Demo)
  YouTube Video

JEDAI: A System for Skill-Aligned Explainable Robot Planning (Demo)

Demo video of JEDAI, winner of the Best Demo Award at AAMAS 2022, showing skill-aligned explainable robot task and motion planning.

AAMAS 2022 (Best Demo Award)  · Mar 9, 2022
Learning User-Interpretable Descriptions of Black-Box AI System Capabilities
  YouTube Video

Learning User-Interpretable Descriptions of Black-Box AI System Capabilities

Talk on discovering high-level, user-interpretable capabilities of black-box planning agents from scratch.

ICAPS 2021 Workshop on Knowledge Engineering for Planning and Scheduling (KEPS)  · Aug 6, 2021
Asking the Right Questions: Active Action-Model Learning
  YouTube Video

Asking the Right Questions: Active Action-Model Learning

Talk on a query-based approach for estimating an interpretable, relational model of a black-box autonomous agent that can plan and act.

AAAI 2021 Workshop on Explainable Agency in Artificial Intelligence  · Feb 2, 2021