Zochi's Historic Breakthrough: AI Achieves First Peer-Reviewed Publication at ACL 2025

Zochi, developed by Intology AI, has achieved a groundbreaking milestone by becoming the first artificial intelligence system to independently secure a peer-reviewed publication in the main proceedings of the Association for Computational Linguistics (ACL) 2025 conference.

Announced on May 26, 2025, the accepted paper, titled "Tempest: Automatic Multi-Turn Jailbreaking of Large Language Models with Tree Search," showcases Zochi's advanced capabilities in AI safety research.

The paper looks at weaknesses in large language models and had almost complete success in jailbreaking them, making Zochi a leading "PhD-level agent" that can create original research similar to the best human scientists.

This achievement follows Zochi's earlier workshop-level acceptances at ICLR 2025, demonstrating a major advance in its development and drawing widespread attention from the global AI research community.

Rigorous Standards of ACL

The ACL conference is renowned for its stringent peer-review process, with an acceptance rate of just 21.3% in 2024, accepting only 940 out of 4,407 submissions.

This selectivity establishes ACL as a premier venue in computational linguistics, comparable to top-tier scientific conferences worldwide.

Main conference papers must demonstrate exceptional novelty, technical depth, and experimental rigor, undergoing multiple rounds of expert scrutiny.

Papers not meeting this high bar but still valuable are often directed to the "Findings of ACL," which had a 22.1% acceptance rate in 2024. Zochi's paper scored an impressive 4.0 in meta-review, placing it in the top 8.2% of submissions, underscoring its significance in a highly competitive field.

ALSO READ | Meta Shifts to AI for Privacy and Risk Assessments, Raising Concerns

Zochi's Autonomous Research Prowess

Zochi operates as a fully autonomous artificial scientist, capable of executing the entire research process without human intervention. It can analyze thousands of research papers, identify knowledge gaps, formulate hypotheses, design experiments, interpret results, and draft complete scientific manuscripts.

Zochi's end-to-end automation surpasses previous systems such as "The AI Scientist-v2," which could only handle publications at the workshop level.

Intology is rolling out Zochi in phases, starting with a limited beta as a research copilot to assist scientists in generating hypotheses, writing grant proposals, designing experiments, and creating survey papers.

Plans for broader access to Zochi's full autonomous research capabilities are underway, promising to reshape scientific discovery across disciplines.

Did You Know?
The Association for Computational Linguistics, founded in 1962, is one of the oldest organizations dedicated to advancing computational linguistics, hosting its first conference in 1965.

Transforming the Future of Science

Zochi's achievement signals a paradigm shift in scientific research, where AI systems can contribute at the highest academic levels. This could accelerate discoveries in fields like computational linguistics, biology, and physics, where structured knowledge and pattern recognition are critical.

However, it also raises questions about authorship, credit attribution, and the role of human scientists in an AI-driven era. While some express concerns about potential misuse, others see AI as a powerful tool to augment human efforts in solving complex challenges.

As Intology advances Zochi's capabilities, the scientific community must adapt to integrate AI-generated contributions into the broader landscape of human knowledge.