Level 1 — Absolute Beginner
Many people work at big companies that build artificial intelligence, or AI. AI is computer technology that can think and learn like a person.
More than 1,200 workers at these companies wrote a letter together. The letter asks the US government for help.
They want the government to build tools that can slow down AI if it grows too fast. They are not asking to stop AI today.
Two big companies, OpenAI and Anthropic, agreed with the letter very quickly. They said AI is already helping build newer AI.
- employee
- a person who works for a company
- artificial intelligence
- computer technology that can perform tasks that normally need human thinking
- letter
- a written message sent to share information or a request
- government
- the group of people who run and make rules for a country
- tool
- something used to help do a job or complete a task
- slow down
- to make something happen at a lower speed
- company
- a business that makes or sells goods or services
- agreed
- said yes to something or shared the same opinion
Level 2 — Elementary
More than 1,200 employees at OpenAI, Anthropic, Google DeepMind and Meta signed an open letter titled 'Pacing the Frontier,' calling on the US government to prepare tools for slowing advanced AI development.
The letter does not ask companies to stop building AI right now. Instead, it asks Washington to help create the technical and governance systems needed for a controlled, verifiable slowdown if one becomes necessary.
Within hours of the letter's release, both OpenAI and Anthropic issued statements endorsing it at the company level, a rare show of agreement between rival AI labs.
Signatories, including Anthropic CEO Dario Amodei and OpenAI chief scientist Jakub Pachocki, warned that leading companies may be close to automating AI research itself, meaning AI systems could soon help design their own successors.
- pacing
- controlling the speed at which something develops
- frontier
- the furthest edge of progress in a field, especially science or technology
- governance
- the systems and rules used to control or manage something
- verifiable
- able to be checked and confirmed as true
- endorsing
- publicly supporting or approving of something
- rival
- a competitor working toward the same goal
- signatory
- a person who signs a document, such as a letter or agreement
- automating
- using machines or software to perform a task without human effort
Level 3 — Intermediate
More than 1,290 employees across OpenAI, Anthropic, Google DeepMind and Meta signed an open letter titled 'Pacing the Frontier,' urging the US government to help develop the technical and governance infrastructure needed to deliberately pace advanced AI development.
The letter is careful to frame its request as preparation rather than prohibition: it explicitly states that nobody is asking for anything to be switched off today, but rather that verifiable mechanisms should exist in case AI capabilities ever outstrip humanity's capacity for safe oversight.
Within hours of publication, both OpenAI and Anthropic took the unusual step of endorsing the letter at the corporate level, a notable alignment between two labs that frequently compete for talent, funding and market share.
Prominent signatories, including Anthropic CEO Dario Amodei, OpenAI chief scientist Jakub Pachocki and Ilya Sutskever, cited internal data suggesting AI now performs a substantial portion of the work that produces subsequent AI generations, a concern Anthropic detailed in its June report 'When AI Builds Itself.'
- infrastructure
- the underlying systems and structures needed to support an activity
- prohibition
- an official act of forbidding something
- mechanism
- a system or process by which something is achieved
- outstrip
- to become greater or faster than something else
- oversight
- the responsibility of supervising and monitoring something
- corporate
- relating to a company as a whole, rather than individuals within it
- alignment
- agreement or cooperation between different parties
- subsequent
- coming after something else in time or order
Level 4 — Advanced
More than 1,290 employees spanning OpenAI, Anthropic, Google DeepMind and Meta have signed an open letter titled 'Pacing the Frontier,' urging Washington to underwrite an international effort to construct the technical and governance infrastructure required to deliberately regulate the tempo of frontier AI development.
The letter is deliberately calibrated to forestall accusations of alarmism, explicitly disclaiming any demand that systems be switched off today, and instead advocating for the advance construction of verifiable braking mechanisms should AI capability growth ever outstrip humanity's capacity for meaningful oversight.
Within hours of publication, OpenAI and Anthropic each issued corporate-level endorsements, an alignment rendered striking by the labs' customary rivalry over talent, capital and market position, and one that lends the letter an institutional weight individual signatures alone could not command.
Prominent signatories, among them Anthropic chief executive Dario Amodei, OpenAI chief scientist Jakub Pachocki and Ilya Sutskever, invoked internal findings indicating that AI now contributes materially to the research pipeline that yields its own successors, a trajectory Anthropic had already substantiated in its June report 'When AI Builds Itself,' which framed recursive self-improvement as a concrete rather than speculative risk.
- underwrite
- to provide financial or organizational support for something
- calibrated
- carefully adjusted for a precise purpose or effect
- forestall
- to prevent something by acting in advance
- alarmism
- the excessive or unwarranted raising of alarm
- disclaiming
- denying responsibility for or connection with something
- institutional
- relating to an established organization rather than an individual
- invoked
- cited or referred to as evidence or justification
- recursive self-improvement
- a process in which a system improves itself, then uses that improvement to improve itself further
Talk about this article with KatePractice this story's vocabulary in a real spoken conversation with your AI tutor