Skip to content

Microsoft president on an AI “kill switch”

Read the Portuguese translation →

Author: Bloomberg Television · Watch on YouTube

Video authored by the channel shown, not by IANIA; all rights belong to the creator. Transcript and summary generated automatically; the original-language content is preserved.

Original language: English

Original video: Microsoft president on an AI “kill switch” · Channel: Bloomberg Television · 00:30 · Watch on YouTube

Topics: AI governance and regulation, AI safety

The speaker argues that AI safety should be built in through multiple layers, rather than relying on a single government-controlled on-off switch. They compare this approach to road safety, which developed through a range of measures such as stoplights, guardrails, and speed limits. The speaker notes that only the speedometer directly changed the car, while the other safeguards supported safety across the broader system. They conclude that a robust, healthy ecosystem and a shared safety framework are essential.

  • AI safety should be built in through multiple layers.
  • A government on-off switch may be useful, but it should not be the only safeguard.
  • The speaker compares AI safety to the development of road-safety measures.
  • Road safety includes stoplights, stop signs, better pavement, guardrails, speed limits, police, and car speedometers.
  • The speedometer was the only item in the list that changed the car itself.
  • Safety requires looking at the ecosystem as a whole.
  • The speaker calls for a robust, healthy ecosystem and a shared safety framework.
Start Chapter Summary
00:00 Layered AI safety The speaker says safety should be built in at multiple levels and questions why an on-off switch should exist in only one place.
00:30 Lessons from road safety Using road-safety measures as an analogy, the speaker argues for addressing safety across the whole ecosystem.
00:30 A shared safety framework The speaker concludes that a robust and healthy ecosystem should come together to establish the safety framework needed.

You really want safety to be built in in multiple layers. — 00:00

why would you put that switch in only one place when you can put it in three or four? — 00:00

the only thing that changed the car was the speedometer. — 00:30

you almost always need to look at the ecosystem as a whole — 00:30

Microsoft, kill switch, safety framework, AI ecosystem

00:00 You really want safety to be built in in multiple layers. At times you may want the government to be able to have the on-off switch. Um but why would you put that switch in only one place when you can put it in three or four? The analog The The analogy I would offer is think about what happened when cars reached the point where they were going fast enough that there were starting to be collisions in an intersection. So, what did people do? They said, “Well, wait a second. We need

00:30 stoplights, stop signs, better pavement on the roads, guardrails, speed limits, a police force, and a speedometer in the car.” But of that list I just had, the only thing that changed the car was the speedometer. Whenever things are going too fast, you almost always need to look at the ecosystem as a whole, ensure it’s robust and healthy, and bring the ecosystem together to put in place the safety framework that’s needed.