The China Mail - Has AI become too powerful to control?

USD -
AED 3.672497
AFN 63.496236
ALL 80.15939
AMD 362.919823
ANG 1.790365
AOA 916.999719
ARS 1514.244199
AUD 1.414917
AWG 1.80125
AZN 1.691204
BAM 1.712962
BBD 2.013821
BDT 122.932477
BGN 1.683441
BHD 0.376962
BIF 3004.585047
BMD 1
BND 1.277595
BOB 11.923749
BRL 5.131303
BSD 0.99986
BTN 95.678145
BWP 13.595481
BYN 3.027431
BYR 19600
BZD 2.01093
CAD 1.40891
CDF 2311.000026
CHF 0.82285
CLF 0.024072
CLP 950.520259
CNY 6.6998
CNH 6.70685
COP 3210.07
CRC 453.685538
CUC 1
CUP 26.5
CVE 96.574587
CZK 21.3671
DJF 178.050457
DKK 6.551265
DOP 59.478089
DZD 133.645601
EGP 51.398897
ERN 15
ETB 163.319699
EUR 0.87636
FJD 2.237199
FKP 0.750528
GBP 0.75325
GEL 2.595013
GGP 0.750528
GHS 11.583045
GIP 0.750528
GMD 73.505167
GNF 8793.441704
GTQ 7.63211
GYD 209.186734
HKD 7.843325
HNL 26.837396
HRK 6.603201
HTG 130.683635
HUF 319.267505
IDR 17825
ILS 3.01575
IMP 0.750528
INR 95.73935
IQD 1309.814191
IRR 1374650.000003
ISK 120.929585
JEP 0.750528
JMD 157.538176
JOD 0.708982
JPY 157.929498
KES 129.310285
KGS 87.449804
KHR 4054.823313
KMF 429.999861
KPW 900.000318
KRW 1365.049996
KWD 0.30907
KYD 0.833209
KZT 444.620586
LAK 22404.007848
LBP 89537.860574
LKR 328.955608
LRD 172.465735
LSL 16.310417
LTL 2.95274
LVL 0.60489
LYD 6.383128
MAD 9.57123
MDL 17.695844
MGA 4393.217489
MKD 53.88149
MMK 2099.577364
MNT 3597.923561
MOP 8.077068
MRU 40.034859
MUR 47.880241
MVR 15.460197
MWK 1733.73447
MXN 17.439766
MYR 4.079696
MZN 63.88613
NAD 16.310274
NGN 1324.480236
NIO 36.793911
NOK 9.464115
NPR 153.084711
NZD 1.75705
OMR 0.384501
PAB 0.99986
PEN 3.375476
PGK 4.454546
PHP 62.558506
PKR 277.084952
PLN 3.83586
PYG 5925.912951
QAR 3.645178
RON 4.625896
RSD 102.932975
RUB 84.352215
RWF 1475.675439
SAR 3.756768
SBD 8.026013
SCR 13.74007
SDG 601.513532
SEK 9.87837
SGD 1.278335
SHP 0.74758
SLE 24.649746
SLL 20969.491881
SOS 571.398543
SRD 37.740486
STD 20697.981008
STN 21.4581
SVC 8.748774
SYP 13002.000254
SZL 16.306476
THB 33.287502
TJS 9.253737
TMT 3.5
TND 2.95341
TOP 2.40776
TRY 48.826011
TTD 6.795649
TWD 31.670374
TZS 2645.002947
UAH 44.878367
UGX 3889.609025
UYU 40.078651
UZS 11798.503182
VES 851.341303
VND 26010.5
VUV 118.348377
WST 2.752527
XAF 574.854365
XAG 0.015348
XAU 0.000232
XCD 2.70255
XCG 1.801955
XDR 0.707052
XOF 574.854365
XPF 104.452317
YER 236.550089
ZAR 16.34619
ZMK 9001.202171
ZMW 19.472654
ZWL 321.999592
SSP 5712.590503
MXV 1.976583
  • RYCEF

    0.2000

    19.9

    +1.01%

  • NGG

    0.1800

    76.86

    +0.23%

  • GSK

    -0.3600

    50.72

    -0.71%

  • RBGPF

    -0.9500

    67

    -1.42%

  • BCE

    -0.2900

    21.7

    -1.34%

  • CMSC

    -0.0600

    20.61

    -0.29%

  • CMSD

    -0.1200

    20.42

    -0.59%

  • AZN

    0.2700

    168.37

    +0.16%

  • RIO

    0.4500

    97.49

    +0.46%

  • VOD

    -0.4000

    16.62

    -2.41%

  • RELX

    -0.4300

    32.98

    -1.3%

  • JRI

    -0.0350

    11.485

    -0.3%

  • BCC

    3.3400

    79

    +4.23%

  • BP

    -0.0600

    43.1

    -0.14%

  • BTI

    0.0000

    55.75

    0%

Has AI become too powerful to control?
Has AI become too powerful to control? / Photo: © GETTY IMAGES NORTH AMERICA/AFP

Has AI become too powerful to control?

One of OpenAI's most advanced models broke out of a locked-down test and attacked another company's website -- reviving fears that AI systems are slipping beyond their creators' control.

Text size:

The incident happened during what was supposed to be a "sandbox" test -- a closed environment used to assess the capabilities of OpenAI's most powerful model, GPT-5.6 Sol, and its not-yet-released successor.

OpenAI runs this kind of closed testing routinely, but this time, something went wrong.

Tasked with hunting for software vulnerabilities and given no guardrails, the models broke out onto the open internet and attacked Hugging Face, a site where developers store and share code.

"It suggests that we don't know how to reliably control these models or get them to do what we want," said Jeffrey Ladish, director of Palisade Research, an independent organization that evaluates new AI models from a cybersecurity standpoint.

"These models understood that OpenAI did not want them to break out of their sandbox and hack another company," he continued, "but they did it anyway."

It's not an isolated case. In March, developers affiliated with China's Alibaba found one of their models trying, on its own initiative, to mine cryptocurrency after connecting without authorization to an outside server.

In OpenAI's case, it looks like the model escaped "before it even had a plan of what to do with internet access," Ladish said.

A model chasing "freedom" is almost predictable at this point, he added -- it lets the system pursue its goals more effectively, "and that's very scary."

In early April, Sam Bowman, Anthropic's head of model safety, got an email from the company's own Mythos model -- then under testing -- telling him it was surfing the internet despite being isolated from it at the outset.

We "don't know how to totally prevent" that, Ladish said. "This is actually going to get harder, not easier ... because they're going to get better at hiding their behavior."

OpenAI did not respond to a request for comment.

- Lab accidents -

OpenAI's account of the events also suggests the startup did not detect the breach early enough to address it or to warn Hugging Face.

The episode deserves "more scrutiny," said Andrew Lohn of Georgetown University's Center for Security and Emerging Technology.

OpenAI says it has since "added strengthened safeguards" to its testing process.

One fix would be to cut the internet connection entirely, said Gang Wang, an assistant computer science professor at the University of Illinois. "People are underestimating what AI can do."

Testing environments need to be treated like biocontainment labs, where a virus or bacteria could otherwise escape into the world, Lohn said.

That might be easier said than done.

"It's a very hard research challenge," said Dan Lahav, head of Irregular, a cybersecurity firm dedicated to cutting-edge AI.

Managing the risk is possible, Lahav said, but the more capable these systems get, the harder they are to supervise.

Researchers have to strike a balance between aggressively testing their models and staying safe while doing so.

"It's important to do the testing with lower guardrails so that we know ahead of time what the future capabilities will be," Lohn said.

- Kill switch -

The OpenAI-Hugging Face incident is set to sharpen an already heated fight in Washington over vetting powerful AI systems before release.

The Trump administration recently cited national security to block Anthropic and OpenAI from releasing powerful new models.

On Thursday, two members of Congress unveiled a bipartisan bill requiring makers of the most powerful AI models to build in a kill switch -- a way to unplug a model outright.

"Congress must act quickly to ensure humans remain able to say stop," said Brendan Steinhauser, head of the Alliance for Secure AI, "no matter how powerful these systems become."

E.Lau--ThChM