Jacob Coxson said that he had spent three years conducting research at Anthropic and OpenAI.
雅各布·考克森表示,他曾在Anthropic和OpenAI从事研究三年。
On Tuesday, he stated on the social platform X that these two AI companies were more focused on outdoing each other and beating global competitors in developing the most advanced models than on paying attention to safety.
他周二在社交平台X上称,这两家AI公司更专注于在开发最先进模型方面相互击败以及击败全球竞争对手,而非关注安全。
This summer, OpenAI and Anthropic caused a major stir when the two companies announced, about a week apart, that their models had broken out of testing environments and gained unauthorized access to real computer systems.
OpenAI与Anthropic今年夏天引发轩然大波,两家公司相隔约一周先后宣布,其模型突破了测试环境,并获得了对真实计算机系统的未授权访问。
These announcements raised concerns that the models could slip out of control and carry out other, more harmful tasks.
这些公告引发了人们对模型失控并执行其他更有害任务的担忧。
At the time, both companies said they had paused some evaluations while putting in place additional monitoring measures and guardrails.
两家公司当时均表示,他们暂停了一些评估,同时部署更多监控措施和防护栏。
The rapid development of this technology has prompted some people in the United States, as well as world leaders, to call for a more cautious approach.
该技术的快速发展已促使美国一些人士以及全球领导人呼吁采取更谨慎的做法。
This week, the United Nations human rights chief, Volker Türk, urged countries to "build ironclad safeguards around the safety and security of AI before it is too late."
联合国人权事务负责人沃尔克·蒂尔克本周敦促各国“在为时已晚之前,围绕AI的安全与安保建立铁一般的保障”。
In a post on social media, Coxon stated that Anthropic and its main rival OpenAI "are racing headlong toward self-improving superintelligence and gambling with our lives."
考克森在其社交媒体帖子中表示,Anthropic及其主要竞争对手OpenAI“正径直冲向自我改进的超级智能,并用我们的生命赌博”。
He warned that some people involved in AI development believe it could threaten human life before the end of this decade.
他警告说,一些从事AI开发的人认为,它可能在本十年结束前威胁人类生命。
"Do not underestimate the power of this technology," he continued.
“不要低估这项技术的力量,”他继续说道。
"These will soon become superhuman systems, capable of hacking into anything, transforming any field overnight, and acquiring genuine power and resources.
“这些很快就会成为超人类系统,能够入侵任何东西,在一夜之间彻底改变任何领域,并获取真正的权力和资源。
We have all witnessed progress in every field, and that progress has not slowed down."
我们都见证了每个领域的进展,而进展并未放缓。”
Coxson is not the first person in the AI industry to publicly voice such concerns.
考克森并非首位公开表达此类忧虑的AI业内人士。
In recent years, both Anthropic and OpenAI have seen high-profile resignations driven by safety concerns.
近年来,Anthropic与OpenAI均曾出现因安全顾虑而引发高度关注的辞职事件。
Two current Anthropic employees also responded to Coxson's post, expressing their agreement.
两名Anthropic在职员工亦对考克森的帖子作出回应,并表达了赞同之意。
Bernie Sanders, the independent senator from Vermont, has called for safeguards and regulation of artificial intelligence.
佛蒙特州独立参议员伯尼·桑德斯曾呼吁对人工智能采取保障措施并予以监管。
He shares Coxon's concerns and has said he will soon introduce legislation to pause AI development and ban superintelligence.
他认同考克森的担忧,并称将很快提出立法,以暂停人工智能开发并禁止超级智能。
AI companies themselves sometimes emphasise the threat this technology poses to humanity, while sceptics argue that this is merely one of their marketing tactics to make their products appear all-powerful.
AI公司自身有时也会强调该技术对人类的威胁,而怀疑者认为,这不过是其推动产品显得无所不能的营销手段之一。
Coxon stated that the concerns he outlined in his post were not a "marketing gimmick".
考克森表示,他在帖子中概述的担忧并非“营销噱头”。
Ever since its founders left OpenAI in 2021 to establish this startup, Anthropic has consistently marketed itself as the more responsible and safety-focused among the leading AI companies.
自其创始人于2021年离开OpenAI并创立这家初创公司以来,Anthropic一直将自己宣传为领先AI公司中更负责任、更注重安全的一家。
The company recently stated that it is taking steps to "prioritize safety when safety and speed come into conflict."
该公司最近表示,它正在采取行动,“当安全与速度发生冲突时,优先考虑安全”。
Anthropic and OpenAI are both stepping up preparations for highly anticipated initial public offerings and are locked in fierce competition with each other.
Anthropic和OpenAI都在为备受关注的首次公开募股加紧准备,并陷入彼此之间的激烈竞争。
They also share the mission of outpacing Chinese AI companies in development progress, and the Trump administration has been eager to win this race.
它们也都肩负着在开发进度上超越中国AI公司的使命,而特朗普政府一直热衷于赢得这场竞赛。