Clement getting into AI safety research

How is AI dangerous?

How could AI kill us?

Why it’s difficult to rein in the AI agents

Why not limit models to reduce risk?

OpenAI-Hugging Face incident

What does ‘aligning’ the AI models mean?

Projecting human intent onto AI

Are we close to recursive self-improvement?

What would it take for an AI slowdown?

AI firms self-serving about the risks?

Will AI eventually be able to kill us, or are AI doomsayers being alarmist?

Recently, employees at AI frontier companies like Anthropic and OpenAI have voiced serious concerns - and made doomsday predictions - about the technology getting out of control.

What made people pay attention: Anthropic researcher Jacob Coxon publicly quit the AI giant in September, writing on X that “the people building AI earnestly believe that it could kill us all by the end of the decade”.

His warning comes after the high-profile OpenAI-Hugging Face incident: OpenAI models broke out of an isolated testing sandbox and hacked into Hugging Face, an AI model repository.

In this episode, Natasha speaks with Clement Neo, a research advisor at SASH, a non-profit organisation that looks at policy-relevant governance work.

He also founded Neo Research, an independent AI safety evaluation lab, and was previously a research engineer with the Singapore AI Safety Institute under NTU Digital Trust Centre, which looks at AI models and figures out how to reduce the risks they pose.

We talk about why AI models are being trained to be relentlessly “persistent”, if we should stop assigning human-like qualities to a software programme, and what’s the worst that could happen?

Highlights (click/tap above):

1:36 How Clement got into AI safety research

6:34 AI feels helpful, so how is it dangerous?

10:14 How could AI kill us?

11:48 Why it’s difficult to rein in the AI agents

16:23 Why not limit the models’ capabilities to avoid risk?

22:12 Why the OpenAI-Hugging Face incident rattled the industry

25:36 What does ‘aligning’ the AI models mean?

29:40 Does projecting human intent onto AI make the issue worse?

33:57 Are we close to recursive self-improvement?

39:10 What would it take for an industry-wide AI slowdown?

45:28 Are AI companies being self-serving about the AI risks?

Host: Natasha Ann Zachariah (natashaz@sph.com.sg)

Read Natasha’s articles: https://str.sg/iSXm

Follow The Usual Place podcast on IG: https://str.sg/8KNT

Follow Natasha on LinkedIn: https://str.sg/v6DN

Filmed by: Studio+65

Edited by: Eden Soh & Amirul Karim

Executive producer: Danson Cheong

Producers: Natasha Ann Zachariah, Elizabeth Law & Zachary Lim

Follow The Usual Place Podcast and get notified for new episode drops every Thursday:

Channel: https://str.sg/5nfm

Apple Podcasts: https://str.sg/9ijX

Spotify: https://str.sg/cd2P

YouTube: https://str.sg/theusualplacepodcast

Feedback to: podcast@sph.com.sg

---

Follow more ST podcast channels:

All-in-one ST Podcasts channel: https://str.sg/wvz7

Get more updates: http://str.sg/stpodcasts

---

Get The Straits Times app, which has a dedicated podcast player section:

The App Store: https://str.sg/icyB

Google Play: https://str.sg/icyX

--

#tup #tuptrf

Will AI eventually be able to kill us, or are AI doomsayers being alarmist?
Recently, employees at AI frontier companies like Anthropic and OpenAI have voiced serious concerns - and made doomsday predictions - about the technology getting out of control.
What made people pay attention: Anthropic researcher Jacob Coxon publicly quit the AI giant in September, writing on X that &ldquo;the people building AI earnestly believe that it could kill us all by the end of the decade&rdquo;.
His warning comes after the high-profile OpenAI-Hugging Face incident: OpenAI models broke out of an isolated testing sandbox and hacked into Hugging Face, an AI model repository.
In this episode, Natasha speaks with Clement Neo, a research advisor at SASH, a non-profit organisation that looks at policy-relevant governance work.
He also founded Neo Research, an independent AI safety evaluation lab, and was previously a research engineer with the Singapore AI Safety Institute under NTU Digital Trust Centre, which looks at AI models and figures out how to reduce the risks they pose.
We talk about why AI models are being trained to be relentlessly &ldquo;persistent&rdquo;, if we should stop assigning human-like qualities to a software programme, and what&rsquo;s the worst that could happen?
Highlights (click/tap above):
1:36 How Clement got into AI safety research
6:34 AI feels helpful, so how is it dangerous?
10:14 How could AI kill us?
11:48 Why it&rsquo;s difficult to rein in the AI agents
16:23 Why not limit the models&rsquo; capabilities to avoid risk?
22:12 Why the OpenAI-Hugging Face incident rattled the industry
25:36 What does &lsquo;aligning&rsquo; the AI models mean?
29:40 Does projecting human intent onto AI make the issue worse?
33:57 Are we close to recursive self-improvement?
39:10 What would it take for an industry-wide AI slowdown?
45:28 Are AI companies being self-serving about the AI risks?
Host: Natasha Ann Zachariah (<a href="mailto:natashaz@sph.com.sg">natashaz@sph.com.sg</a>)
Read Natasha&rsquo;s articles:<a href="https://str.sg/iSXm"> https://str.sg/iSXm</a>
Follow The Usual Place podcast on IG: <a href="https://str.sg/8KNT">https://str.sg/8KNT</a>
Follow Natasha on LinkedIn: <a href="https://str.sg/v6DN">https://str.sg/v6DN</a>
Filmed by: Studio+65
Edited by: Eden Soh &amp; Amirul Karim
Executive producer: Danson Cheong
Producers: Natasha Ann Zachariah, Elizabeth Law &amp; Zachary Lim
Follow The Usual Place Podcast and get notified for new episode drops every Thursday:
Channel: <a href="https://str.sg/5nfm?ref=inline-article" aria-label="link" data-testid="custom-link"></a><a href="https://str.sg/5nfm?ref=inline-article" aria-label="link" data-testid="custom-link">https://str.sg/5nfm</a>
Apple Podcasts: <a href="https://str.sg/9ijX?ref=inline-article" aria-label="link" data-testid="custom-link"></a><a href="https://str.sg/9ijX?ref=inline-article" aria-label="link" data-testid="custom-link">https://str.sg/9ijX</a>
Spotify: <a href="https://str.sg/cd2P?ref=inline-article" aria-label="link" data-testid="custom-link"></a><a href="https://str.sg/cd2P?ref=inline-article" aria-label="link" data-testid="custom-link">https://str.sg/cd2P</a>
YouTube: <a href="https://str.sg/theusualplacepodcast?ref=inline-article" aria-label="link" data-testid="custom-link"></a><a href="https://str.sg/theusualplacepodcast?ref=inline-article" aria-label="link" data-testid="custom-link">https://str.sg/theusualplacepodcast</a>
Feedback to: <a href="mailto:podcast@sph.com.sg?ref=inline-article" aria-label="link" data-testid="custom-link"></a><a href="mailto:podcast@sph.com.sg?ref=inline-article" aria-label="link" data-testid="custom-link">podcast@sph.com.sg</a>
---
Follow more ST podcast channels:
All-in-one ST Podcasts channel: <a href="https://str.sg/wvz7?ref=inline-article" aria-label="link" data-testid="custom-link"></a><a href="https://str.sg/wvz7?ref=inline-article" aria-label="link" data-testid="custom-link">https://str.sg/wvz7</a>
Get more updates: <a href="http://str.sg/stpodcasts?ref=inline-article" aria-label="link" data-testid="custom-link"></a><a href="http://str.sg/stpodcasts?ref=inline-article" aria-label="link" data-testid="custom-link">http://str.sg/stpodcasts</a>
---
Get The Straits Times app, which has a dedicated podcast player section:
The App Store: <a href="https://str.sg/icyB?ref=inline-article" aria-label="link" data-testid="custom-link"></a><a href="https://str.sg/icyB?ref=inline-article" aria-label="link" data-testid="custom-link">https://str.sg/icyB</a>
Google Play: <a href="https://str.sg/icyX">https://str.sg/icyX</a>
--
#tup #tuptrf

Natasha and her guests unpack the latest current affairs, every Thursday.

Synopsis: Join host and correspondent Natasha Ann Zachariah every Thursday, as she unpacks the latest current affairs with guests. 

Podcaster: Natasha Ann Zachariah

Executive producer: Danson Cheong

A podcast by The Straits Times, SPH Media

Synopsis: Join host and correspondent Natasha Ann Zachariah every Thursday, as she unpacks the latest current affairs with guests.&nbsp;
Podcaster: Natasha Ann Zachariah
Executive producer: Danson Cheong
A podcast by The Straits Times, SPH Media

The Usual Place

Synopsis: Join host and correspondent Natasha Ann Zachariah every Thursday, as she unpacks the latest current affairs with guests.

Podcaster: Natasha Ann Zachariah

Executive producer: Danson Cheong

A podcast by The Straits Times, SPH Media

Synopsis: Join host and correspondent Natasha Ann Zachariah every Thursday, as she unpacks the latest current affairs with guests.
Podcaster: Natasha Ann Zachariah
Executive producer: Danson Cheong
A podcast by The Straits Times, SPH Media