Newshot

奥尔特曼 says OpenAI can't satisfy everyone and must balance AI safety with development

AI summary

At the DealBook summit on Wednesday, November 18, OpenAI CEO 萨姆 · 奥尔特曼 (Sam Altman) said the company is seeking a balance between "safety" and "development." 奥尔特曼 said OpenAI cannot "satisfy everyone," warning that racing to pursue breakthroughs would fuel public fears of AI getting out of control, while overemphasizing safety could slow progress, and he stressed the company's "iterative deployment" strategy of rolling out technology in stages and observing risks in real-world settings.

Translation(model translation, not the source's own text)

Article title: 奥尔特曼 says OpenAI can't please everyone and must balance AI safety with development

Body:

  • Day/Night Follow system Light Dark

奥尔特曼 says OpenAI can't please everyone and must balance AI safety with development

IT Home, November 19 news: At the DealBook summit held on Wednesday, November 18, US local time, OpenAI CEO 萨姆 · 奥尔特曼 (Sam Altman) said that his company is seeking a balance between "safety" and "development" in AI.

奥尔特曼 and OpenAI President 格雷格 · 布罗克曼 (Greg Brockman) both took part in the event. They said that risks are unavoidable in the process of advancing AI, so the company must make trade-offs between prioritizing safety and continuing to innovate.

In the interview, 奥尔特曼 acknowledged that OpenAI cannot "please everyone." He said that if the company only "floors the gas" in pursuit of technological breakthroughs, the public will worry about AI running out of control; but if it overemphasizes safety, it may also slow product and technological progress.

奥尔特曼 stressed that OpenAI adopts an "iterative deployment" strategy, that is, rolling out technology in stages, observing risks in real environments and continuously improving. In his view, one important way to understand AI risks cannot be entirely separated from reality through closed-door development; more importantly, technology must be tested by reality.

奥尔特曼 also admitted that as model capabilities rapidly increase, safety, alignment and monitoring are becoming significantly more difficult. OpenAI President and co-founder 格雷格 · 布罗克曼 said that as model capabilities have increased significantly, safety, security and "alignment" (that is, AI behavior conforming to human intent) have become issues that must be taken seriously. He said this is not just a shift for OpenAI alone, but a cultural shift the entire AI industry is undergoing.

OpenAI has signed voluntary agreements with other AI companies, committing to conduct risk assessments and cooperate with third-party evaluators. The company said this week that it is bringing in independent safety researchers from organizations such as METR and Apollo Research to conduct safety evaluations of its models.

奥尔特曼 also warned that the current fierce competition among AI companies is not conducive to safety. He said each company can only understand its own safety standards and finds it difficult to fully grasp competitors' development; this "race in the fog" may lead to insufficient safety investment. He called for the industry to establish common safety standards and promote greater transparency.

Regarding some discussions in Silicon Valley about AI gradually taking over decision-making power and even possibly forming "superintelligence," 奥尔特曼 explicitly opposed it. He said humans must continue to control the future and should not gradually hand decision-making power to AI; 布罗克曼 also believes that the stronger AI capabilities become, the greater the responsibility to ensure they are aligned with human goals and remain safe and controllable.

On the business side, 奥尔特曼 revealed that OpenAI still plans to go public at some point. Previously, OpenAI postponed its IPO plan to next year to prioritize safety issues; earlier this month, the company completed 11 billion US dollars (IT Home note: about 73.489 billion yuan at current exchange rates) in financing, with a valuation of 1.25 trillion US dollars (about 8.35 trillion yuan at current exchange rates).

Related articles

  • OpenAI CEO 奥尔特曼 warns: in the future, humanity may be unable to take back control from AI
  • OpenAI 奥尔特曼 on memory shortage: global DRAM should prioritize data centers, idle in gamers' PCs is waste
  • 奥尔特曼 responds to AI impact: understands mathematicians' situation, but OpenAI will not stop research
  • OpenAI 奥尔特曼: in the face of AI, mathematicians initially feel "we're doomed," but ultimately accept reality
  • OpenAI CEO 奥尔特曼 accused of sexual abuse by his sister, case dismissed by judge
  • 英伟达 黄仁勋 publicly rebuts OpenAI 奥尔特曼's view: must have "zero tolerance" for AI cybersecurity

Sites under Ruimei: IT Home Zuihuimai - rebate and cashback coupons iPhone Home Win7 Home Win10 Home Win11 Home

Software under Ruimei: Ruimei Mobile APP Applications Magic Cube Zuihuimai Yaozhi

The translation is machine-generated and individual terms may be off. Nothing is rewritten or cut; the original is authoritative.

Published by IT之家
This site only summarises and aggregates; copyright stays with the original author. Headlines and summaries follow the source — where they differ, the original wins.
Read the original ↗