People who build AI

Ray Kurzweil

inventor and author of The Singularity Is Near

3 records

Records

Artificial Intelligence is the most difficult to contend with, because whereas we can articulate technical solutions to the dangers of biotech, there‘s no purely technical solution to a so-called unfriendly AI. We can‘t just say, “We‘ll just put this little software code sub-routine in our AIs, and that‘ll keep them safe.” I mean, it really comes down to what the goals and intentions of that artificial intelligence are.

Ray Kurzweil

inventor and author of The Singularity Is Near

In a 2009 interview, Ray Kurzweil said unfriendly AI had no purely technical solution and that its safety depended on its goals and intentions.

Inventor and futuristPublished December 30, 2009 · date of remarks not establishedh+ magazine interview, archived on Ray Kurzweil's official website

Source Archived

Context and checks

Surrounding words

We‘ll be overcoming problems at a very rapid rate, and that will seem magical. But that doesn‘t mean it‘s not rooted in science and technology. I say it‘s not utopian because it also introduces new problems. Artificial Intelligence is the most difficult to contend with, because whereas we can articulate technical solutions to the dangers of biotech, there‘s no purely technical solution to a so-called unfriendly AI. We can‘t just say, “We‘ll just put this little software code sub-routine in our AIs, and that‘ll keep them safe.” I mean, it really comes down to what the goals and intentions of that artificial intelligence are. We face daunting challenges. RUS: I think when most people think of utopia, they probably just think about everybody being happy and feeling good. RK: I really don‘t think that‘s the goal.

What this quote does not say

The original h+ magazine URL failed to open, and the official archive describes its page as a summary rather than a complete transcript.
Earlier in the same interview, Kurzweil said he was fairly optimistic humanity would make it through existential risks and that painful episodes were more likely.

How this was checked

Quote matched character-for-character against the captured page.

Reviewed by

an AI reviewer that read the whole source · an automatic character-by-character check

Overall, we should be cautiously optimistic. While AI is creating new technical threats, it will also radically enhance our ability to deal with those threats. As for abuse, since these methods will enhance our intelligence regardless of our values, they can be used for both promise and peril. We should thus work toward a world where the powers of AI are broadly distributed, so that its effects reflect the values of humanity as a whole.

Ray Kurzweil

inventor and author of The Singularity Is Near

In a 2024 op-ed, Ray Kurzweil called for cautious optimism, saying AI creates new technical threats but can help address them and its power should be spread widely.

Author of The Singularity is Nearer: When We Merge With AIPublished June 25, 2024 · date of remarks not establishedThe Promise and Peril of AI

Source Archived

Context and checks

Surrounding words

We are not doomed to failure in controlling these perils. Overall, we should be cautiously optimistic. While AI is creating new technical threats, it will also radically enhance our ability to deal with those threats. As for abuse, since these methods will enhance our intelligence regardless of our values, they can be used for both promise and peril. We should thus work toward a world where the powers of AI are broadly distributed, so that its effects reflect the values of humanity as a whole.

What this quote does not say

This is a qualification, not a reversal: Kurzweil still affirms technical threats, abuse, and peril.
The passage gives no probability estimate and does not distinguish catastrophic risk from lower-severity technical threats.

How this was checked

Quote matched character-for-character against the captured page.

Reviewed by

an AI reviewer that read the whole source · an automatic character-by-character check

Regarding the Open Letter to “pause” research on AI “more powerful than GPT-4,” this criterion is too vague to be practical. And the proposal faces a serious coordination problem: those that agree to a pause may fall far behind corporations or nations that disagree. There are tremendous benefits to advancing AI in critical fields such as medicine and health, education, pursuit of renewable energy sources to replace fossil fuels, and scores of other fields.

Exact quote, beginning

Ray Kurzweil

inventor and author of The Singularity Is Near

In April 2023, Ray Kurzweil said he did not sign a proposed six-month pause on advanced AI research because its standard was vague and called for more tailored safety measures.

Inventor, best-selling author, and futuristApril 2023My opinion on the request for 6 month delay on large language models that go beyond GPT-4

Source Archived

Context and checks

Surrounding words

April 22, 2023 by Ray Kurzweil April 2023 Regarding the Open Letter to “pause” research on AI “more powerful than GPT-4,” this criterion is too vague to be practical. And the proposal faces a serious coordination problem: those that agree to a pause may fall far behind corporations or nations that disagree. There are tremendous benefits to advancing AI in critical fields such as medicine and health, education, pursuit of renewable energy sources to replace fossil fuels, and scores of other fields. I didn’t sign, because I believe we can address the signers’ safety concerns in a more tailored way that doesn’t compromise these vital lines of research. I participated in the Asilomar AI Principles Conference in 2017 and was actively involved in the creation of guidelines to create Artificial Intelligence in an ethical manner. So I know that safety is a critical issue.

What this quote does not say

He does not specify what the preferred tailored safety measures are in this passage.
This rejects one pause proposal; it does not reject all regulation, pauses, or safety constraints.

How this was checked

Quote matched character-for-character against the captured page.

Reviewed by

an AI reviewer that read the whole source · an automatic character-by-character check