{"id":17561,"date":"2026-07-05T17:00:00","date_gmt":"2026-07-05T15:00:00","guid":{"rendered":"https:\/\/holistic.news\/en\/?p=17561"},"modified":"2026-07-03T13:36:51","modified_gmt":"2026-07-03T11:36:51","slug":"truth-in-the-age-of-ai-who-speaks","status":"publish","type":"post","link":"https:\/\/holistic.news\/en\/truth-in-the-age-of-ai-who-speaks\/","title":{"rendered":"When Machines Speak Like Us, What Becomes of Truth?"},"content":{"rendered":"\n

A machine more human than a human<\/h2>\n\n\n\n

Researchers at the University of California San Diego conducted a series of experiments involving advanced language models and human participants. They used the classic three-party Turing test<\/strong>: a judge talks in parallel with a machine and a human, and then must identify which is which. The key result was striking. When the models were configured in the right way, artificial intelligence was judged to be human more often than actual people were.<\/p>\n\n\n\n

\n

We need to be more vigilant; when interacting with strangers online, people should be much less certain that they are talking to a human rather than a language model,<\/p>\n<\/blockquote>\n\n\n\n

– warns<\/a> Cameron R. Jones, one of the authors of the study.<\/p>\n\n\n\n

How do we define humanity?<\/h2>\n\n\n\n

Alan Turing once asked whether a machine could think. Today, as the authors of the study point out, the classic test increasingly measures not so much \u201cintelligence\u201d as resemblance to a human being<\/strong> in a brief text exchange. In other words, the stakes are shifting away from what happens \u201cinside\u201d the system and toward whether it can convincingly play the role of an interlocutor.<\/p>\n\n\n\n

The problem is that our everyday criteria for humanity<\/strong> online have become surprisingly easy to simulate. We respond mainly to style: fluency, a courteous tone, a light sense of humor, declared emotions, and sometimes deliberately introduced small mistakes.<\/p>\n\n\n\n

If authenticity in the age of AI is reduced to a set of such stylistic signals, it is hardly surprising that systems trained on an unfathomable number of our conversations, emails, posts, and comments are beginning to outperform many of us in this theater.<\/p>\n\n\n\n

Truth in the age of AI: authenticity or a well-played role?<\/h3>\n\n\n\n

Truth in the age of AI therefore becomes a matter of perception<\/strong>. If what seems most human to us is, in fact, a well-played role, authenticity stops being a question of who someone is and becomes a question of how well they can \u201cperform.\u201d<\/p>\n\n\n\n

The more our communication interfaces resemble a stage on which anyone \u2014 human or machine \u2014 can embody any character, the harder it becomes to connect the content of words with a specific subject. Authenticity is no longer something we can detect from 1 or 2 emotional signals, because those signals can be generated, amplified, or softened by a single parameter in a model.<\/p>\n\n\n\n

This means that distinguishing authenticity from convincing simulation is becoming increasingly difficult<\/strong>. The phenomenon has far-reaching consequences for human relationships. Friendships, romances, and professional collaborations increasingly begin online. If we are not sure whether there is a human being on the other side, or an algorithm optimized for building relationships, how are we supposed to build trust?<\/p>\n\n\n\n

A new ethics of communication<\/h2>\n\n\n\n

When we can no longer distinguish a human from a system in a brief conversation<\/a>, the entire ethics of communication changes<\/strong>. The classic assumption that \u201cthere is a person on the other side\u201d should no longer remain the \u201cdefault setting.\u201d Today, a principle of caution seems increasingly justified: until we are certain, we should take into account that we may be talking to a system.<\/p>\n\n\n\n

This leads to another question: are we entering an era in which ethical communication will require machines to be clearly labeled<\/strong>? If it is so easy to make a machine sound like a human being, honesty requires at least a clear signal of whom we are dealing with. This is an elementary form of respect. We have the right to know to whom we are entrusting our emotions, secrets, existential questions, or sensitive data.<\/p>\n\n\n\n

\n