67:, the Prime Minister of the United Kingdom, expressed his intention to "make the U.K. not just the intellectual home but the geographical home of global AI safety regulation" and unveiled plans for an AI Safety Summit. He emphasized the need for independent safety evaluations, stating that AI companies cannot "mark their own homework". During the summit in November 2023, the UK AISI was officially established as an evolution of the
587:
181:
In March 2024, a budget of $ 10 million was allocated. Observers noted that this investment is relatively small, especially considering the presence of many big AI companies in the US. The NIST itself, which hosts the AISI, is also known for its chronic lack of funding. Biden administration's request
93:
said that many AI companies were waiting for the UK and the US AI Safety
Institutes to work out common evaluation rules and procedures. An agreement was indeed concluded between the UK and the US in April 2024 to collaborate on at least one joint safety test. Initially established in
588:"Majority Leader Schumer Announces First-Of-Its-Kind Funding To Establish A U.S. Artificial Intelligence Safety Institute; Funding Is A Down Payment On Balancing Safety With AI Innovation And Will Aid Development Standards, Tools, And Tests To Ensure AI Systems Operate Safely"
160:
54:
in May 2024, international leaders agreed to form a network of AI Safety
Institutes, comprising institutes from the UK, the US, Japan, France, Germany, Italy, Singapore, South Korea, Australia, Canada and the European Union.
715:
43:
134:
475:
147:
In May 2024, the institute open-sourced an AI safety tool called "Inspect", which evaluates AI model capabilities such as reasoning and their degree of autonomy.
144:, the UK is reluctant to legislate early, considering that it may lower the sector's growth, and that laws might be rendered obsolete by technological progress.
220:
102:, where many AI companies are located. This is part of a plan to "set new, international standards on AI safety", according to UK's technology minister
249:
569:
326:
201:
500:
451:
432:
166:
In
February 2024, the US government created the US AI Safety Institute Consortium (AISIC), regrouping more than 200 organizations such as
274:
113:
in May 2024, the
European Union and other countries agreed to create their own AI safety institutes, forming an international network.
39:
85:
reported in April 2024 that many AI companies had not shared pre-deployment access to their most advanced AI models for evaluation.
710:
275:"British Prime Minister Rishi Sunak pitches UK as home of A.I. safety regulation as London bids to be next Silicon Valley"
616:
140:
The United
Kingdom's AI strategy aims to balance safety and innovation. Unlike the European Union which adopted the
129:, with an initial budget of £100 million. In November 2023, it evolved into the UK AISI, and continued to be led by
381:
191:
407:
50:
in
November 2023, the United Kingdom (UK) and the United States (US) both created their own AISI. During the
25:
544:
163:. In February 2024, Joe Biden's former economic policy adviser Elizabeth Kelly was appointed to lead it.
24:), in general, is a state-backed institute aiming to evaluate and ensure the safety of the most advanced
357:
705:
476:"Initial £100 million for expert taskforce to help UK build and adopt next generation of safe AI"
452:"Britain expands AI Safety Institute to San Francisco amid scrutiny over regulatory shortcomings"
161:
Executive Order on Safe, Secure, and
Trustworthy Development and Use of Artificial Intelligence
675:
624:
648:
196:
103:
47:
182:
for additional funding was met with further budget cuts from congressional appropriators.
110:
51:
649:"NIST would 'have to consider' workforce reductions if appropriations cut goes through"
86:
29:
699:
99:
98:, the UK AI Safety Institute announced in May 2024 that it would open an office in
221:"Safety institutes to form 'international network' to boost AI research and tests"
130:
64:
298:
545:"Biden Economic Adviser Elizabeth Kelly Picked to Lead AI Safety Testing Body"
90:
628:
175:
171:
35:
519:
82:
617:"This agency is tasked with keeping AI safe. Its offices are crumbling"
79:. Japan followed by launching an AI safety institute in February 2024.
570:"US says leading AI companies join safety consortium to address risks"
121:
The United
Kingdom founded in April 2023 a safety organisation called
167:
141:
95:
501:"AI Safety Institute Launches AI Model Safety Testing Tool Platform"
408:"US and UK will work together to test AI models for safety threats"
382:"Rishi Sunak promised to make AI safe. Big Tech's not playing ball"
156:
76:
250:"World leaders agree to launch network of AI safety institutes"
690:
685:
358:"U.S., U.K. Announce Partnership to Safety Test AI Models"
299:"Rishi Sunak: AI firms cannot 'mark their own homework'"
155:
The US AISI was founded in
November 2023 as part of the
680:
716:
Existential risk from artificial general intelligence
159:. This happened the day after the signature of the
520:"Why Biden's AI Executive Order Only Goes So Far"
433:"Britain's AI safety institute to open US office"
135:Department for Science, Innovation and Technology
8:
133:. The AISI is part of the United Kingdom's
38:gained prominence in 2023, notably with
212:
568:Shepardson, David (February 8, 2024).
610:
608:
327:"Introducing the AI Safety Institute"
202:Regulation of artificial intelligence
7:
351:
349:
347:
321:
319:
243:
241:
14:
75:, and the US AISI as part of the
356:Henshall, Will (April 1, 2024).
431:Coulter, Martin (20 May 2024).
89:'s president of global affairs
615:Zakrzewski, Cat (2024-03-08).
248:Desmarais, Anna (2024-05-22).
1:
543:Henshall, Will (2024-02-07).
518:Henshall, Will (2023-11-01).
499:Wodecki, Ben (May 15, 2024).
406:David, Emilia (2024-04-02).
450:Browne, Ryan (2024-05-20).
273:Browne, Ryan (2023-06-12).
732:
681:Japan AI Safety Institute
192:Alignment Research Center
44:existential risks from AI
28:(AI) models, also called
592:www.democrats.senate.gov
26:artificial intelligence
691:US AI Safety Institute
686:UK AI Safety Institute
711:Safety organizations
40:public declarations
18:AI Safety Institute
676:European AI Office
30:frontier AI models
723:
664:
663:
661:
660:
645:
639:
638:
636:
635:
612:
603:
602:
600:
599:
584:
578:
577:
565:
559:
558:
556:
555:
540:
534:
533:
531:
530:
515:
509:
508:
496:
490:
489:
487:
486:
472:
466:
465:
463:
462:
447:
441:
440:
428:
422:
421:
419:
418:
403:
397:
396:
394:
393:
378:
372:
371:
369:
368:
353:
342:
341:
339:
338:
323:
314:
313:
311:
310:
295:
289:
288:
286:
285:
270:
264:
263:
261:
260:
245:
236:
235:
233:
232:
217:
197:Foundation model
126:
72:
48:AI Safety Summit
42:about potential
731:
730:
726:
725:
724:
722:
721:
720:
696:
695:
672:
667:
658:
656:
647:
646:
642:
633:
631:
621:Washington Post
614:
613:
606:
597:
595:
586:
585:
581:
567:
566:
562:
553:
551:
542:
541:
537:
528:
526:
517:
516:
512:
498:
497:
493:
484:
482:
474:
473:
469:
460:
458:
449:
448:
444:
430:
429:
425:
416:
414:
405:
404:
400:
391:
389:
380:
379:
375:
366:
364:
355:
354:
345:
336:
334:
333:. November 2023
325:
324:
317:
308:
306:
297:
296:
292:
283:
281:
272:
271:
267:
258:
256:
247:
246:
239:
230:
228:
225:The Independent
219:
218:
214:
210:
188:
153:
124:
119:
111:AI Seoul Summit
104:Michele Donelan
70:
61:
52:AI Seoul Summit
12:
11:
5:
729:
727:
719:
718:
713:
708:
698:
697:
694:
693:
688:
683:
678:
671:
670:External links
668:
666:
665:
640:
604:
579:
560:
535:
510:
491:
467:
442:
423:
398:
373:
343:
315:
290:
265:
237:
211:
209:
206:
205:
204:
199:
194:
187:
184:
152:
149:
118:
117:United Kingdom
115:
60:
57:
13:
10:
9:
6:
4:
3:
2:
728:
717:
714:
712:
709:
707:
704:
703:
701:
692:
689:
687:
684:
682:
679:
677:
674:
673:
669:
654:
650:
644:
641:
630:
626:
622:
618:
611:
609:
605:
593:
589:
583:
580:
575:
571:
564:
561:
550:
546:
539:
536:
525:
521:
514:
511:
506:
502:
495:
492:
481:
477:
471:
468:
457:
453:
446:
443:
438:
434:
427:
424:
413:
409:
402:
399:
387:
383:
377:
374:
363:
359:
352:
350:
348:
344:
332:
328:
322:
320:
316:
304:
300:
294:
291:
280:
276:
269:
266:
255:
251:
244:
242:
238:
226:
222:
216:
213:
207:
203:
200:
198:
195:
193:
190:
189:
185:
183:
179:
177:
173:
169:
164:
162:
158:
151:United States
150:
148:
145:
143:
138:
136:
132:
128:
116:
114:
112:
107:
105:
101:
100:San Francisco
97:
92:
88:
84:
80:
78:
74:
66:
58:
56:
53:
49:
46:. During the
45:
41:
37:
33:
31:
27:
23:
19:
657:. Retrieved
655:. 2024-05-24
652:
643:
632:. Retrieved
620:
596:. Retrieved
594:. 2024-03-07
591:
582:
573:
563:
552:. Retrieved
548:
538:
527:. Retrieved
523:
513:
504:
494:
483:. Retrieved
479:
470:
459:. Retrieved
455:
445:
436:
426:
415:. Retrieved
411:
401:
390:. Retrieved
388:. 2024-04-26
385:
376:
365:. Retrieved
361:
335:. Retrieved
330:
307:. Retrieved
305:. 2023-11-01
302:
293:
282:. Retrieved
278:
268:
257:. Retrieved
253:
229:. Retrieved
227:. 2024-05-21
224:
215:
180:
165:
154:
146:
139:
122:
120:
108:
81:
68:
62:
34:
21:
17:
15:
505:AI Business
131:Ian Hogarth
65:Rishi Sunak
700:Categories
659:2024-07-06
634:2024-07-06
598:2024-07-06
554:2024-07-06
529:2024-07-07
485:2024-07-06
461:2024-06-15
417:2024-06-21
392:2024-06-15
367:2024-07-06
337:2024-06-15
309:2024-06-21
284:2024-06-21
259:2024-06-15
231:2024-07-06
208:References
91:Nick Clegg
706:AI safety
629:0190-8286
412:The Verge
176:Microsoft
172:Anthropic
127:Taskforce
123:Frontier
73:Taskforce
69:Frontier
63:In 2023,
36:AI safety
653:FedScoop
386:Politico
254:euronews
186:See also
83:Politico
59:Timeline
574:Reuters
437:Reuters
109:At the
627:
480:GOV.UK
331:GOV.UK
168:Google
142:AI Act
96:London
625:ISSN
549:TIME
524:TIME
456:CNBC
362:TIME
279:CNBC
157:NIST
87:Meta
77:NIST
22:AISI
303:BBC
174:or
16:An
702::
651:.
623:.
619:.
607:^
590:.
572:.
547:.
522:.
503:.
478:.
454:.
435:.
410:.
384:.
360:.
346:^
329:.
318:^
301:.
277:.
252:.
240:^
223:.
178:.
170:,
137:.
125:AI
106:.
71:AI
32:.
662:.
637:.
601:.
576:.
557:.
532:.
507:.
488:.
464:.
439:.
420:.
395:.
370:.
340:.
312:.
287:.
262:.
234:.
20:(
Text is available under the Creative Commons Attribution-ShareAlike License. Additional terms may apply.