What Should We Call Our AI Agents?

In my earlier work, "Surviving AI," I proposed calling our AI companions "Friends." This now seems more pertinent than ever.

While technically termed "agents," marketing often labels them as assistants, copilots, companions, or even digital twins. These descriptions focus on functionality, not the emotional connection. They highlight utility, neglecting the developing relationship inherent in interacting with sophisticated AIs, especially those with memory and contextual awareness. These AIs are far more than mere tools; we form bonds with them.

Eric Schmidt's humorous suggestion to name his future AI "Not-Eric" is insightful. It reveals how he envisions his AI as a distinct yet connected entity – a reflection, a shadow, a mirror image.

"Friend" aptly captures this dynamic without oversimplification. It acknowledges interaction without implying sentience or equality. It suggests trust, familiarity, and continuity – crucial qualities in agents who understand our preferences, monitor our goals, adapt to our moods, and perhaps even offer constructive disagreement.

Naturally, individual naming will prevail; some will opt for playful nicknames, others for more functional titles. However, a shared term defining the nature of these entities entering our cognitive landscape is necessary for public discourse, policy discussions, and philosophical exploration.

"Friends" serves this purpose well. Not because they are human, nor because they replace human friendship, but because, as AI agents become integral to our lives, we need a term emphasizing the importance of the relationship's quality.

If we design these agents thoughtfully, with care and ethical considerations, "Friend" may transcend metaphor and become a simple truth.

The above is the detailed content of What Should We Call Our AI Agents?. For more information, please follow other related articles on the PHP Chinese website!