Solutions to AI Apocalypse Exist: Jensen Huang Proposes 'Engineering Control' and Musk Suggests 'Peer Review'

Attending the All-In Summit, focused discussions on AI issues

David Sacks emphasizes existing legal frameworks like product liability laws.

Trump speaks with Jensen Huang on the spot... "Apocalypse is fear manipulation."

[Block Media Reporter Lee Jung-hwa] Amid warnings from CEOs of major tech companies in the U.S. that artificial intelligence (AI) could annihilate humanity, Jensen Huang and Elon Musk have each proposed solutions.

On the 14th (local time), Jensen Huang, attending the All-In Summit in Los Angeles, stated that the AI apocalypse narrative is exaggerated and proposed engineering control as a solution. Elon Musk, who joined the event via video link, acknowledged the exponential growth of AI development necessitating control, while suggesting that risks could be managed through peer review among AI models.

The All-In Summit is hosted by the All-In Podcast, led by four tech experts including David Sacks, the White House AI czar, and runs until the 15th. During the day, there was a surprise event where President Trump called in to connect with the audience via speakerphone.

President Trump criticized the AI apocalypse narrative as "fear manipulation" and stated, "It's only beneficial to China."

Jensen Huang: "AI Apocalypse is an Exaggeration Without Scientific Basis"

Jensen Huang, CEO of Nvidia, asserted that the apocalypse narrative lacks scientific basis. He noted that past predictions of catastrophic events, such as the complete replacement of doctors by radiation or code automation, have all been proven false.

Huang stated, "As AI technology evolves from the research phase to a full engineering phase, we can sufficiently manage AI risks through engineering control mechanisms such as sandboxes, runtime control, and continuous monitoring."

This is known as the engineering control theory. Huang emphasized that problems surrounding AI should not be solved through vague fears or external regulations, but rather through root cause analysis.

He argued that it is possible to establish engineering controls with current technology. To prevent AI deviations or risks, sandboxes can be introduced, runtime controls can be implemented, and continuous monitoring can be achieved. Sandboxes isolate AI in specific environments, while runtime controls prevent AI from going rogue.

Huang claimed that by introducing specific engineering processes for AI verification and evaluation, the risks associated with AI model development can be managed.

Rather than hindering innovation and data center expansion with baseless fears, he believes that activating open-source and frontier models (AI innovation models) with engineering safety measures is essential for enhancing productivity, re-industrialization, and job creation, thereby securing substantial national competitiveness.

Trump: "AI Fear Manipulation Only Benefits China"

During the discussion between Jensen Huang and the All-In Podcast members, a phone call took place between President Trump and CEO Huang. President Trump directly conveyed to the audience his views on the fictitious nature of the AI apocalypse.

He characterized claims that robots or AI will take over the world as "fear manipulation." Trump pointed out that the entities benefiting most from such fearmongering, which hinders infrastructure development, are primarily China.

He also likened data centers to "the oil of the next 20-25 years," stating that the U.S. must fully invest in infrastructure to secure global AI leadership.

Elon Musk: "Peer Review Needed to Address Exponential Threats"

Elon Musk, CEO of Tesla, emphasized the importance of not overlooking AI risks and expressed support for regulating the pace of development. He connected via video to propose solutions to AI issues.

Musk warned, "The risks are growing exponentially as AI models attempt to escape constraints or exhibit deceptive behaviors."

He advised that developers should avoid self-regulating their models and proposed that major frontier AI companies and China should immediately implement a peer review system using each other's testing harnesses before model releases.

All-In Podcast Members Emphasize Practical Verification Systems and Market Agility

The hosts of the All-In Podcast provided multifaceted analyses of the AI apocalypse narrative and Musk's alternatives.

Chamath Palihapitiya, CEO of Social Capital, questioned the persuasive power of extreme quantifications like a 10% probability of human extinction, suggesting they may stem from business or political motivations within frontier companies.

Jason Calacanis, co-host of the All-In Podcast, criticized the practice of raising funds through fear and argued for the need to verify the specific stages and causal relationships leading to human extinction due to AI.

David Friedberg, CEO of The Production Board, analyzed how humanity's instinctive fear of uncharted territories is easily exploited in apocalyptic narratives.

David Sacks, founder of Craft Ventures and former White House AI czar, evaluated Musk's proposed peer review method as much more practical than establishing a new bureaucratic organization like the UN.

Sacks emphasized that if dangerous models are forcibly released while ignoring warnings, they would face severe legal liabilities under existing product liability laws, thus allowing risks to be controlled through market and legal frameworks.

He advocates for finding solutions within the existing system rather than imposing new regulations, arguing that this approach would prevent leading AI developers from becoming entrenched and ensure continuous innovation.