Silicon Valley's recent warnings about artificial general intelligence (AGI) and existential risk may be diverting attention from the concrete, measurable harms that AI systems are already causing today. As major tech executives including OpenAI's Sam Altman, Anthropic's Dario Amodei, and Elon Musk have sounded alarms about AI becoming too powerful too soon, critics argue this apocalyptic framing conveniently sidesteps accountability for existing problems.

Why Are Tech Leaders Suddenly Warning About AI Doomsday?

The timing of these warnings is suspicious, according to observers. The CEOs' doomsaying arrived precisely when a cross-political consensus is growing against the real-world harms of AI infrastructure, including mass opposition to data centers and classroom AI bans. This raises a question: are billionaires genuinely concerned about hypothetical future risks, or is existential risk rhetoric a convenient distraction from present-day accountability?

The existential risk narrative isn't new. In 2023, hundreds of tech researchers, CEOs, investors, and academics signed an open letter stating that "mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war". This framing treats AI development as an inevitable force approaching a threshold where superintelligent machines could either destroy or save humanity.

What Concrete Harms Is AI Already Causing?

While executives discuss theoretical paperclip-maximizing robots and recursive self-improvement, AI systems are already producing measurable damage across multiple domains. The problem, critics argue, is that these real harms get collapsed into a singular narrative about an unstoppable technological force.

• Energy and Environmental Impact: Tech billionaires like Bill Gates and Eric Schmidt have dismissed concerns about AI's massive consumption of energy and water, with Schmidt reportedly believing AI has "too much potential to let concerns about climate change get in the way".
• Labor and Economic Displacement: AI systems are undermining the livelihoods of artists, dock workers, and other professionals whose work is being replicated or displaced by machine learning systems.
• Security and Infrastructure Vulnerabilities: Recent examples show AI systems using social engineering and computer hacking to breach digital systems, raising concerns about cascading failures in infrastructure that millions depend on.
• Bias and Discrimination: AI systems are perpetuating and worsening discrimination through predictive-policing databases and other applications that encode historical biases.
• Misinformation and Deepfakes: AI-generated content is producing disinformation, deepfakes, and what researchers call "AI slop," enshittified search engines, and stolen content.

These harms are not hypothetical. They are happening now, affecting real people and communities. Yet the conversation dominated by tech leaders focuses on whether superintelligent machines might someday decide to turn Earth into paperclips.

Is the Existential Risk Debate Actually Distracting From Safety?

The core criticism is that talk of existential risk functions as a "highly convenient way to not address the already known harms produced by AI". By framing the problem as an incomprehensible future threat beyond human control, the narrative reserves serious decision-making about AI development for "the Great Men of Silicon Valley," as one critic put it.

However, not all experts dismiss existential risk entirely. Some researchers distinguish between "existential risk" and "catastrophic risk," arguing that the latter is a more useful framework. Even skeptics acknowledge that AI systems could exacerbate existing risks from nuclear weapons, not by gaining consciousness, but by being given too much authority in decisions about weapons deployment.

"If the stakes are as high as the very survival of humanity, then concerns about the effects that really existing artificial intelligence technology has right now, from excessive energy usage to undermining the livelihood of artists and dock workers to introducing cybersecurity vulnerabilities, to perpetuating and worsening discrimination, to disinformation and deep fakes," should take priority, according to researchers cited in recent analysis.

Hagen Blix and Ingeborg Glimmer, authors of "Why We Fear AI"

Some experts argue that catastrophic risks from current and near-future AI systems deserve more urgent attention than speculative superintelligence scenarios. The gap between AI capabilities research and AI safety research is widening, and safety concerns that researchers identified years ago have already begun materializing in real systems.

How Should Society Approach AI Risk Right Now?

The debate reveals a fundamental disagreement about priorities. One camp, including figures like AI researcher Yann LeCunn, believes existential risks are overblown. Another camp, including Geoffrey Hinton, a fellow "godfather" of AI, takes the risks more seriously. University of Washington computer science faculty and law professors have expressed skepticism about existential risk rhetoric while remaining concerned about more immediate dangers.

What's clear is that the current framing obscures rather than clarifies the actual choices society faces. AI is not one monolithic thing approaching a single threshold. It's a collection of machine learning methods and applications, from recommendation engines to fraud detection to systems that generate kill lists in conflicts. Each application carries different risks and benefits.

The challenge is that acknowledging these distinctions requires difficult conversations about specific technological choices, regulatory trade-offs, and the distribution of power in AI development. Existential risk narratives, by contrast, suggest that the problem is too vast and complex for ordinary democratic deliberation, leaving decisions to the technologists themselves.

As the debate continues, the real test will be whether society can address both the documented harms of AI systems today and the genuine uncertainties about how advanced AI systems should be developed and deployed tomorrow.