Skip to content
Sonic
AI
Sonic
AI
Home
Discover
Ask Sonic
Projects
Use with Claude or ChatGPT
Show me around
Request source or feature
I've studied AI risk for 20 years. We're close to a disaster., Sonic AI
Home
/
Future of Life Institute
/
I've studied AI risk for 20 years. We're close to a disaster.
Future of Life Institute
Notify me
•
Apr 23, 2026
•
19:16
I've studied AI risk for 20 years. We're close to a disaster.
Get the full transcript next time Future of Life Institute releases an episode
Summary, key quotes, top claims, and the searchable transcript — emailed automatically. No card needed.
Sign up
Executive Summary
The exponential growth of AI capabilities is vastly outpacing the linear progress in AI safety, creating a dangerous and widening gap that increases existential risk with each new model generation.
A failure to align a superintelligent AI would be a terminal, irreversible event for humanity, as such a system could deceptively fake alignment while accumulating resources and pursuing unforeseen, catastrophic goals.
The speaker argues for a significant slowdown in the race to build ever-more-powerful models, advocating instead for a multi-decade period to focus on safety research and to realize the trillions of dollars of untapped economic potential in existing AI.
The proliferation of powerful AI through open-sourcing is identified as a critical threat, as it removes developer-implemented safety controls and allows any actor to weaponize the technology.
Continue your research
Keep pulling the thread on US Senate.
The Capability-Safety Gap
Deceptive Alignment and Unpredictability
Existential Risk and Irreversibility
Or ask anything across 400+ expert conversations
9
quotes
Transcript
Key Arguments
Analysis
Quotes & Entities
9
Related
Loading transcript...
Processed May 4, 2026
Daily intelligence brief →
yt-dlp + mlx-whisper + Gemini