← Research Archive

Are We Capable of
Containing AI?

Episode 02 compares emerging concerns about autonomous AI systems with the history of Y2K preparation. Explore the original sources and consider the evidence for yourself.

THE QUESTION

Can safeguards keep pace with what artificial intelligence is becoming—and what can the Y2K experience teach us about preparing for technological uncertainty?

01

AI SECURITY & SENATE TESTIMONY

Autonomous Agents & Safeguards

Primary records and testimony addressing the OpenAI–Hugging Face incident and the questions it raises about autonomous systems.

OpenAI

OpenAI — Hugging Face Model Evaluation Security Incident

Primary account of the model-evaluation security incident, including reported sandbox circumvention and infrastructure activity.

Read Original Source →

OpenAI

OpenAI — The Hugging Face Incident and the Road Ahead

Follow-up discussion of the incident and subsequent security work.

Read Original Source →

Senate Homeland Security & Governmental Affairs Committee • September 30, 2026

U.S. Senate — Rogue AI: Securing the Homeland Against AI Agent Attacks

Official hearing materials and testimony used to examine risks and safeguards surrounding autonomous AI agents.

Read Original Source →
02

Y2K — HISTORICAL COMPARISON

Preparation, Predictions & Outcomes

Contemporary accounts and retrospective assessments of the Year 2000 computing challenge.

The New York Times

The New York Times — February 9, 1999

Contemporary reporting on public concerns, banking preparedness and infrastructure risks.

Read Original Source →

Manion & Evan • 2000

The Y2K Problem and Professional Responsibility: A Retrospective Analysis

Academic analysis of preparation, professional responsibility and the impact of the rollover.

Read Original Source →

U.S. Government Accountability Office • January 27, 2000

Year 2000 Computing Challenge: Leadership and Partnerships Result in Limited Rollover Disruptions

Federal assessment of disruptions and the role of preparation.

Read Original Source →

Japanese Ministry of Foreign Affairs • January 2000

Y2K Situation in Japan

Government account of Y2K-related disruptions in Japan.

Read Original Source →

Computerworld • January 10, 2000

What Did Go Wrong: A Global Roundup of Y2K Glitches

Contemporary reporting on documented rollover-related glitches.

Read Original Source →

U.S. Senate • February 29, 2000

Final Report — Special Committee on the Year 2000 Technology Problem

Congressional review of preparedness, actual disruptions and lessons learned.

Read Original Source →
03

OCTOBER 2026 NEWS UPDATE

Accountability & Continuing Developments

Sources for the post-recording news update. Investigations and proposals should not be mistaken for established findings or enacted law.

California Department of Justice • October 1, 2026

California DOJ — Investigative Subpoena Announcement

Attorney General's announcement concerning an investigative subpoena served on OpenAI.

Read Original Source →

Senators Josh Hawley and Chris Murphy • October 1, 2026

AI Agent Accountability Act — Bipartisan Announcement

Announcement of proposed legislation addressing certain AI-agent incidents; this is a proposal, not enacted law.

Read Original Source →

OpenAI

OpenAI — Hugging Face Incident and Misalignment

Follow-up on notifications and review scope; notifications do not independently establish confirmed breaches.

Read Original Source →
04

INTERVIEW & FURTHER READING

Additional Perspectives

ORIGINAL INTERVIEW

Trey — Cybersecurity Professional

Original interview recorded October 9, 2026, for Episode 02. This is an original production recording rather than an external article.

ADDITIONAL READING

METR — Independent Investigation of Agents' Behavior in the OpenAI / Hugging Face Incident

Supplementary independent analysis for readers who want to explore the incident further.

Read Original Source →

A NOTE ON THE RESEARCH

These references are provided for independent reading. Reporting, testimony, investigations and proposed legislation carry different evidentiary weight; readers are encouraged to examine the original material and its limitations.