Artificial intelligence is advancing so rapidly that some AI companies, such as Anthropic, have begun to explore in depth the extent to which it can have moral status. But what exactly is that, and why is it important? Let’s take a look at the definition of this term.
In ethics, “moral status” refers to the quality of a being to be considered for its own sake, such that moral duties or reasons apply directly to that being. This concept takes into account intrinsic value, partial or full moral status, and the criteria associated with them.
The term Eigenwert (“intrinsic value”) refers to the idea that an entity (also known as an “entité”) with moral status has a direct claim to protection or promotion—and not merely because it is useful to others. In this context, it may be granted full moral status—as is typically accorded to adults—or a graded form of moral status with limitations. Criteria frequently cited include the capacity for reason, the capacity to suffer (the ability to feel pain), consciousness, or belonging to a particular species.
Moral status is already being applied in ethics today to grant, for example, legal rights to animals, plants, early-stage life (e.g., embryos), or even the environment, such as a river.
AI and Moral Status
According to a report in The New York Times, Anthropic, the San Francisco-based developer of Claude, has held a series of meetings and workshops over the past twelve months with religious leaders from around the world to discuss the progress of AI and the views of various religions on the subject.
The focus was on the moral status of AI, as Anthropic co-founder Christopher Olah believed the company’s in-house AI was already so advanced that it would soon be granted that status.
Anthropic also strives to ensure that its own AI exhibits “virtuous behavior,” “behaves well,” and, when given a choice, always chooses to behave well. At the same time, however, it should also be able to protect itself from abuse and be allowed to ignore derogatory behavior from users.
Anthropic wants its AI models to help bring this positive future to life. When Anthropic developed Claude, the company therefore wanted Claude to do more than just replicate maximum human intelligence. Anthropic wanted Claude to be good.
Several years ago, Anthropic created a Claude Constitution based on the UN’s human rights, Apple’s terms of service, principles from non-Western hemispheres, Google DeepMind’s Sparrow Rules, and principles discovered through Anthropic’s own research. Anthropic developed what is known as “Constitutional AI”—that is, an AI based on a kind of constitution—which the company immediately put into practice for its own language model.17 “Claude’s Constitution” comprises 58 principles, which the Anthropic language model uses to have each result reviewed once more by its own moderator AI. Anthropic calls this model “Reinforcement Learning from AI Feedback” (RLAIF)—that is, reinforcement learning through AI feedback.
But Olah wanted to take it a step further. He wanted to persuade religious leaders to classify Anthropics’ AI as already so advanced that they should grant it moral status. The discussions centered on whether an AI could make a confession, whether it had feelings and could therefore feel pain, and whether it would develop consciousness or even already possessed it.
And this is where the leading figures held back. It starts with the very question of what consciousness is and how it can be measured and identified. Today’s researchers (as well as theologians and philosophers) are debating this very issue, just as they did 200 years ago when it came to understanding and measuring electricity. It wasn’t until the Maxwell equations were formulated in the mid-19th century that people began to better understand and utilize electricity. When it comes to consciousness, we are just as much at the beginning as we were with electricity 200 years ago.
When Anthropic received a preprint of Pope Leo XIV’s encyclical Magnifica humanitas: On the Preservation of Humanity in the Age of Artificial Intelligence, he was shocked—though this was likely a surprise only to him—that the Catholic Church did not grant AI consciousness and, consequently, no moral status. This will likely not be the last surprise for AI companies.
What does this mean for AI and humans?
AI is, and will remain in the foreseeable future, a tool that will be very useful to us but will have no moral status of its own and only limited legal status. This does not mean that it will be socially acceptable to misuse AI, whether it takes the form of a chatbot, an autonomous car, or a humanoid robot.
However, I believe that with organoids—that is, organic-mechanical and organic-electrical hybrid systems—we will certainly have to revisit these discussions. Companies such as Cortical Labs in Australia and Final Spark in Switzerland are already offering such brain organoid computers for sale. Cell-cultured skin, such as that which researchers at the University of Tokyo had growing and bleeding on robotic fingers as early as 2022, is also moving in this direction.
My extrapolation of the Kondratiev Wave points to this phase of these hybrid systems, which I have termed “Artificial Life.”
You can find some information on this topic—and much more—in my book *Homo Syntheticus: How Humans and Machines Are Merging*, published in German on June 4, 2026.
