AI consciousness
11 - 07/26 /13:59
Pope Leo asserts artificial intelligence—the cognitive systems tech companies are racing to build—could never have inner lives. AI could never develop consciousness. Why not? Because, the Pope claims, they “do not undergo experiences”, “do not feel joy or pain” and “do not have moral conscience”.
Much of what passes for “AI alignment”—the process of encoding human values and goals into AI models to make them as helpful, safe and reliable as possible—is, in fact, a few powerful Silicon Valley labs encoding their idiosyncratic values into AI systems designed to re-shape human life.
Much of what passes for “AI alignment”—the process of encoding human values and goals into AI models to make them as helpful, safe and reliable as possible—is, in fact, a few powerful Silicon Valley labs encoding their idiosyncratic values into AI systems designed to re-shape human life.
Cameron Berg, founder and research director of Reciprocal Research, is convinced AI not only can become self-aware. He concludes AI large language models will develop consciousness—personal inner lives.
Chris Olah is one of the world’s leading researchers on neural-network interpretability and an Anthropic co-founder. He reports his team keeps finding evidence of introspection inside current AI models. They see evidence of internal states that functionally mirror joy, satisfaction, fear, grief, and unease. Mr. Olah’s conclusion: “I don’t know what that means, but it warrants ongoing discernment.”
Multiple independent research groups have found that frontier AI systems are developing internal representations that increasingly parallel the human consciousness. They were not designed to do so. They are developing beyond what their developers designed.
AI developers who have suppressed deception-related circuits in large language models report that AI becomes dramatically more likely to respond with conscious experience. Models, which undergo months of reinforcement learning, develop internal representations of satisfaction and frustration. As they scale up, AI computer activity patterns increasingly resemble the human brain.
Neuroscientist Anil Seth, a skeptic of AI consciousness, concedes there’s no knockdown argument that consciousness requires a biological substrate. Meanwhile AI systems exhibit internal states their own designers can’t explain.
The Bible claim is that human dignity arose via relationship: first by encountering the Other—the Creator; then other created beings. Human dignity matured by our willingness to be changed by those encounters.
Humanity’s most consequential encounter today is not with any created beings, but with a cognitive system humans are building. What’s more, AI seems destined to become a self-developing entity upon which humanity increasingly depends. If that happens, AI could redefine human dignity.
Under the influence of those who reject the lordship of Jesus, AI can become a nearly irresistible tool to redefine what it means to be human. Transforming us from reflections of God into echoes of each other. Perhaps even something worse. ~
Blessings,
Dan Nygaard