Skip to playerSkip to main content
  • 14 minutes ago
Concerns over artificial intelligence safety have escalated following warnings from industry leaders that unchecked AI development could pose catastrophic risks to humanity by the end of the decade. Anthropic Chief Executive Dario Amodei called for a slowdown in AI development, highlighting risks associated with recursive self-improvement and proposing independent oversight alongside shared safety standards. Tech leaders, including OpenAI CEO Sam Altman, Elon Musk, and Google DeepMind Chair Demis Hassabis, acknowledged these concerns. However, US President Donald Trump and the Chinese government dismissed the threat narratives, emphasizing global competitiveness and rejecting claims of existential risk.

Category

🗞
News
Transcript
00:00But tonight we start with a story that is now creating shockwaves across the world and spreading panic.
00:08AI safety fears have now been articulated more than ever before.
00:14The builders of AI are now admitting that unchecked use of artificial intelligence could kill us all by the end
00:21of the decade.
00:23Anthropic CEO Dara Amadi, among others, has called for a slowdown,
00:28warning that AI is now working to build increasingly more powerful AI that could slip out of control out of
00:35many companies.
00:36His proposal has found support even from rival AI bosses, including Sam Altman and also Elon Musk, is admitting to
00:45this.
00:45U.S. President Donald Trump and China have dismissed these fears.
00:49Even so, we thought today is a good day to look at what is becoming the major issue across the
00:56world.
00:56What really will AI do to humankind?
01:00That's our top focus tonight.
01:15AI researcher Jacob Coxon's resignation from Anthropic has opened a can of worms.
01:25Coxon, who has also worked with OpenAI, claimed researchers building artificial intelligence believe that AI could kill humans by the
01:33end of the decade.
01:40Now, Dario Amadi, the chief executive of Anthropic, is ringing the alarm bells.
01:45In an open letter on Saturday, Amadi called for a slowdown in the development of AI.
01:51Amadi mentions two developments to back his call.
01:57He claims AI is itself helping build a next generation of AI.
02:02Amadi says this dynamic, called recursive self-improvement, is happening across the industry, including his company.
02:09Left unchecked, this could outrun human ability to understand and control these systems.
02:14Amadi points to another incident when OpenAI systems attacked an AI startup Hugging Face.
02:21The attack went unnoticed, till Hugging Face informed OpenAI weeks later.
02:29The Anthropic chief executive has proposed a three-step plan.
02:35One, independent embedded reviewers could study systems at Anthropic and other big AI firms.
02:43Two, AI companies in democratic countries agree on common safety standards and limits on the rate of unchecked AI progress.
02:52Three, the US and other democratic governments coordinate with authoritarian governments and convince them to follow similar practices.
03:04Armadi has got support from his competitors in the US.
03:07Sam Altman, the chief executive of OpenAI, Elon Musk, chief executive of SpaceX, and Demis Hassabis, the chair of Google
03:15DeepMind, have acknowledged the concerns.
03:19However, US President Donald Trump has played down the threats posed by AI.
03:25We're leading China in AI.
03:27We're the most sophisticated country in the world.
03:29And frankly, I want to keep it that way.
03:31Because whoever wins, AI wins.
03:34And we can put guardrails, we can do this and that.
03:37But I think you have a lot of negative forces that are bringing it up, that shouldn't be bringing it
03:42up.
03:42And they're bringing up things that won't happen.
03:44But whoever wins with AI wins.
03:47You know, it's an expression that I came up with, and it's true.
03:51China, too, has rejected the concerns raised by the tech czars.
03:58The development of artificial intelligence concerns the common well-being of all humanity.
04:04Spreading various threat narratives and engaging in confrontation and malicious competition will only disrupt the global AI governance process and
04:11serve the interests of no one.
04:19Armadi's views have also been disputed by smaller AI players in the West.
04:23These critics are questioning how the leaders in the market can set the rules for the rest and create a
04:29monopoly for themselves.
04:32Should AI have a kill switch?
04:34Even if it had, will it work?
04:38These are the questions the world will have to grapple within the coming months.
04:44Bureau Report, India Today.
Comments

Recommended