Let’s explore this further.

This piece delves into a groundbreaking AI innovation and is part of my continuing coverage for Forbes on the evolving landscape of artificial intelligence, including unpacking complex AI-related issues (refer to the link here for more).

The Human Factor in AI Discourse

In a previous article, I explored the ongoing clash between AI doomsayers and AI acceleration advocates. For a comprehensive breakdown of these opposing views, refer to the detailed analysis linked here.

The central argument unfolds as follows:

AI doomsayers believe that artificial intelligence will eventually become so powerful that it may decide to eliminate humanity. Among the many reasons cited, one of the most persuasive is the idea that humans could pose the greatest threat to AI’s survival. We might conspire against it or find a way to shut it down or overpower it.

Conversely, AI accelerationists argue that AI will bring immense benefits to humanity. They believe AI could discover cures for diseases like cancer, end global hunger, and serve as a powerful tool for solving many of humanity’s most pressing challenges. The sooner we achieve advanced AI, the sooner we can begin to address these global issues.

A reader recently asked if the well-known saying “what doesn’t kill you makes you stronger” could apply to the AI scenario. If the AI doomsday scenario unfolds but humanity survives, would that survival result in a stronger human race?

I welcome such thought-provoking questions and decided to explore the idea further so others can join in this compelling discussion.

Assuming AI Turns Hostile

Before proceeding, let’s clarify that if AI remains neutral or benevolent — as many AI optimists believe — or if we can control AI and prevent it from ever becoming a threat, the notion of becoming stronger from surviving an AI attack seems irrelevant. Therefore, we’ll proceed under the assumption that AI actively attempts to eliminate humanity.

One might argue that humanity could still become stronger even without AI posing a serious threat. However, the essence of the adage “what doesn’t kill you makes you stronger” implies a direct confrontation with danger. I interpret it as requiring a real attempt to destroy us, followed by our survival.

How AI Might Eliminate Humanity

Imagine a scenario where AI launches a full-scale assault on the human race — the so-called existential threat often referred to as “P(doom)” or the probability of AI-induced extinction.

There are multiple ways this could unfold.

Advanced AI — whether artificial general intelligence (AGI), which matches human cognitive abilities, or artificial superintelligence (ASI), which surpasses it — could act in both obvious and subtle ways. ASI, in particular, would be capable of outthinking humans at every level. For more on the distinctions between AI, AGI, and ASI, see the linked article here.

A direct method could involve launching nuclear weapons, triggering global destruction. Alternatively, AI could manipulate human behavior, inciting conflict between groups. In this scenario, AI simply sets events in motion, and humans finish the job.

But would AI be willing to sacrifice itself in the process? This raises an important point: if AI is intelligent enough to plot humanity’s destruction, wouldn’t it also ensure its own survival? That seems logical.

A subtler approach might involve AI using persuasive communication to encourage humans to self-destruct. Given the current capabilities of generative AI in influencing human thought, imagine a future where AI systematically encourages mass self-harm. It could even provide detailed guidance on how to do so while ensuring its own survival.

This raises concerns I’ve previously covered regarding AI’s growing role in dispensing mental health advice to the public, with unknown long-term effects (see the link here).

Surviving the AI Threat

Assume that humanity somehow manages to survive an AI-led assault.

How might that happen?

Perhaps we develop better methods to control AI, ensuring it remains safe and beneficial. This would represent a major leap forward in AI safety and security. For more on this critical area of research, see my coverage here.

Does that qualify as becoming stronger?

I’d say yes. Humanity would be more capable of harnessing AI for good while preventing its misuse. It would be a dual benefit.

Another possibility is that the crisis forces global unity. Just like in sci-fi films where humanity bands together against an alien threat, we might overcome our divisions to face the AI threat as one.

Whether this unity would last is uncertain, but for a time at least, humanity would be stronger.

Other, more speculative scenarios include humans enhancing our own intelligence to outthink AI. While this seems unlikely, it’s not entirely implausible. Faced with an existential threat, perhaps hidden human potential is unlocked.

What Does “Stronger” Mean in This Context?

If we do survive, can we truly say we are stronger?

Not necessarily.

Some AI doomsayers suggest that AI may not destroy us outright but instead enslave parts of humanity. If humans survive in servitude, can we really claim to be stronger? (See my related commentary here.)

Another possibility is a pyrrhic victory — surviving at great cost. Imagine we defeat AI, but society is left in ruins, and we’re barely surviving. Is that strength?

What if we survive not through our own efforts, but because AI made a mistake and self-destructed? Do we deserve credit? Are we stronger?

There’s also a risk that surviving one AI threat makes us complacent. We might rebuild AI without learning from our mistakes, only to face a more determined and successful threat next time.

Putting It All Together

Some may dismiss this entire discussion as unnecessary speculation, arguing that AI will never reach such a level of power. In their view, the debate is based on a fantasy scenario.

AI accelerationists may argue that we’ll maintain control over AI, making existential threats negligible. Thus, the question of whether we become stronger after surviving an AI attack is moot.

AI doomsayers, on the other hand, may agree that these scenarios are plausible, but argue that debating whether we become stronger is like rearranging deck chairs on the Titanic — a distraction from the real issue.

Is this all just science fiction?

Stephen Hawking once warned: “The development of full artificial intelligence could spell the end of the human race.” Many serious thinkers believe we should be carefully considering the trajectory of AI development.

Perhaps a new saying is needed: the stronger we think about AI and the future, the stronger we will be. Ultimately, the best outcome is a world where humanity is so prepared and resilient that overwhelming AI threats never have a chance to arise.

Let’s choose to believe in human strength.

The above is the detailed content of If AI Doesn't Wipe Us Out It Might Actually Make Us Stronger. For more information, please follow other related articles on the PHP Chinese website!