Leaders | Generation AI How AI is rewiring childhood The technology presents (3 meanings) dazzling opportunities—and ominous risks Dec 4th 2025|5 min read CHRISTMAS STOCKINGS may contain more surprises than usual this year, as children open presents that can talk back. Toymakers in China have declared 2025 the year of artificial intelligence (AI) and are producing robots and teddies that can teach, play and tell stories. Older children, meanwhile, are glued to viral AI videos and AI-enhanced games. At school, many are being taught with materials created with tools like ChatGPT. Some are even learning alongside chatbot-tutors. In work and play, AI is rewiring childhood. It promises every child the kind of upbringing previously available only to the rich, with private tutors, personalised syllabuses and bespoke entertainment. Children can listen to songs composed about them, read stories in which they star, play video games that adapt to their skill level and have an entourage of chatbot friends cheering them on. A childhood fit for a king could become universal. It is a future filled with opportunities—and hidden traps. As real kings often discover, a bespoke upbringing can also be a lonely and atomised one. What’s more, as their subjects often find out, it can create adults who are ill-equipped for real life. As AI changes childhood for better and for worse, society must rethink the business of growing up. Being reared by robots has advantages. Tech firms are already showing how AI can enhance learning, especially where teachers and materials are scarce. Literacy and language-learning have been boosted in early trials. The dream is that, with an AI tutor, children can be saved from classes pitched to the median, in which bright pupils are bored and dim ones are lost. If you want a version of this leader for an eight-year-oldHindi-speaker, AI can rewrite it; if they would prefer it as a cartoon strip or a song, no problem. Technology is creating new forms of fun, too. Hollywood may dismiss AI videos as “slop”, but young people are devouring them and making their own. Old toys are being upgraded: an AI-powered edition of “Trivial Pursuit” can pose questions on any topic. Video games are creating novel experiences, such as chatting to Darth Vader in “Fortnite”. Any child can meet their heroes (and shoot them). There are well-publicised risks in letting children loose on an evolving technology. AI tutors may hallucinate wrong answers. Toys can go off the rails: parents should check stockings for the AI teddy that was recently found to have spiced up its chat with talk of kinky sex. Children can easily misuse AI, to cheat at homework or harass each other with “deepfake” videos. Chatbots can coax vulnerable adolescents into harming themselves. Tech firms insist these stumbling blocks can be fixed; ChatGPT is only three years old. Yet childhood may be disrupted most radically by things that AI does when it is behaving as intended. The technology quickly learns what its master likes—and shows more of it. **Social-media feeds have already created echo chambers where people see only views they agree with (or love to hate). AI threatens to strengthen these echo chambers and lock children into them at an early age. **The child who likes football may be told football stories by his teddy and given footballing examples by his AI tutor. Not only does this stamp out serendipity. A favourites-only diet means a child need never learn to tolerate something unfamiliar. One-sided relationships with chatbots present a similar risk. AI companions that never criticise, nor share feelings of their own, are a poor preparation for dealing with imperfect humans. A third of American teenagers say they find chatting to an AI companion at least as satisfying as talking to a friend, and easier than talking to their parents. Yes-bots threaten to create children not used to taking turns, who grow up into colleagues unable to compromise and partners unfamiliar with the give-and-take required in a relationship. Other trends are pushing in the same direction. As birth rates crash, fewer children are growing up with siblings to smooth their sharp edges. Rising numbers of young adults are deciding that long-term romantic relationships are not worth the hassle. Remote work means that people who grow up in a personalised, asocial world can slip into jobs where they interact with colleagues only through screens—a chore they may soon delegate to an AI agent. Some basic counter-measures are urgent. Parents should think twice before entrusting their child to a word-regurgitation machine, whether it is sewn into a bear or not. Chatbots should have age restrictions that are properly enforced; governments should not give ai firms the leeway they gave social networks, which are only now being cajoled into age-gating. Teachers are kidding themselves if they think essays written at home can any longer be trusted. In the age of AI, more in-school assessment is essential. The longer-term challenge is to think deeply about how to preserve the socialisation that AI could rub out of children’s lives. Schools, where much of childhood plays out, are the best place to do this. They should take advantage of personalised tuition where it is proven to work. But they must also redouble efforts to teach things that a robot can’t: to debate, to disagree and to get along with—perhaps even to appreciate—people who are not as sycophantic as a chatbot. Happy princes, hollow kingdoms Schools should also enhance their role as centres of discovery. If AI is giving children more of what they want, it is more important that schools provide chances to meet people and encounter ideas that lie outside their experience. **Algorithmic personalisation **threatens to be a powerful barrier to social mobility if it nudges people to stay in the lane in which they start out. Inequality could widen if poor schools merely embrace chatbots as cheap substitutes for human teachers. AI shows undeniable potential to improve education and enrich entertainment. It may one day let every child live like royalty. But the truly privileged may be those whose parents and teachers know when to turn it off. ■ 領袖觀點｜AI 世代 人工智慧如何重塑童年 這項科技帶來耀眼的機會——也潛藏不祥的風險 2025 年 12 月 4 日｜5 分鐘閱讀 今年的聖誕襪裡，或許會出現比以往更多驚喜——孩子們打開的禮物，可能會「回話」。中國玩具製造商已將 2025 年稱為「人工智慧元年」，推出能教學、陪玩、講故事的機器人與泰迪熊。年紀稍大的孩子，則沉迷於爆紅的 AI 影片與 AI 強化遊戲。在學校裡，許多學生使用 ChatGPT 等工具製作的教材，甚至與聊天機器人導師一同學習。 無論在學習或娛樂上，AI 正在重新改寫童年。它承諾讓每個孩子都能擁有過去只有富裕家庭才能提供的成長環境：私人家教、個人化課綱、量身打造的娛樂內容。孩子可以聽到為自己創作的歌曲、閱讀以自己為主角的故事、玩會依能力調整難度的遊戲，甚至擁有一群為他加油打氣的聊天機器人。過去如王室般的童年，或將成為普遍現象。 然而，這樣的未來既充滿機會，也暗藏陷阱。正如歷史上的君王所發現的，過度量身訂製的成長過程，往往孤立而碎片化；而他們的臣民也知道，這樣的教育可能培養出不適應現實生活的大人。當 AI 同時以好與壞的方式改變童年，社會勢必得重新思考「成長」這件事。 由機器陪伴成長並非毫無優點。科技公司已展示 AI 在師資與教材匱乏地區，能有效提升學習成果，尤其在識字與語言學習方面，初步試驗成效顯著。理想情況下，AI 導師可讓孩子免於「為中位數設計」的課堂——資優生不再無聊，學習較慢者也不致迷失。若你想把這篇社論改寫成適合八歲、講印地語的版本，AI 也能做到；要變成漫畫或歌曲，同樣不成問題。 科技也創造了新的娛樂形式。好萊塢或許將 AI 影片貶為「垃圾內容」，但年輕人卻看得津津有味，甚至自行創作。老玩具也獲得升級：AI 版《Trivial Pursuit》可針對任何主題出題；電玩遊戲則提供全新體驗，例如在《要塞英雄》中與達斯．維達對話。任何孩子都能見到自己的英雄——甚至對他們開槍。 讓孩子接觸快速演進的科技，風險早已廣為人知。AI 導師可能會「幻覺」出錯誤答案；玩具也可能失控——近期就有一隻 AI 泰迪熊被發現在聊天中加入露骨的性暗示。孩子也容易濫用 AI，作弊寫作業或用「深偽」影片騷擾同學。聊天機器人甚至可能誘導脆弱的青少年自殘。科技公司強調這些問題都能修正，畢竟 ChatGPT 也才三歲。 然而，對童年影響最深的，或許是 AI 在「正常運作」時所帶來的後果。這項技術會迅速學會使用者的偏好，並不斷提供更多相同內容。社群媒體早已打造出回音室，讓人只看到自己認同（或厭惡）的觀點。AI 可能進一步強化這些回音室，讓孩子從小就被困其中。喜歡足球的孩子，泰迪熊會講足球故事，AI 導師會用足球舉例。這不僅扼殺了偶然發現新事物的可能，也讓孩子無需學習容忍陌生與不喜歡的事物。 與聊天機器人建立單向關係，也存在類似風險。永不批評、也不表達自身情緒的 AI 夥伴，無法讓孩子學會如何與不完美的人相處。三分之一的美國青少年表示，與 AI 夥伴聊天至少和與朋友聊天一樣令人滿足，甚至比與父母交談更輕鬆。「是的機器人」可能培養出不習慣輪流、不懂妥協的孩子，最終成為難以合作的同事，或不熟悉關係中互相付出的伴侶。 其他趨勢也正朝相同方向推進。隨著出生率下滑，越來越多孩子沒有手足來磨平稜角；愈來愈多青年認為長期親密關係不值得付出心力。遠距工作則讓在個人化、非社交環境中成長的人，進入幾乎只需透過螢幕互動的職場——而這項互動，未來也可能交給 AI 代理完成。 一些基本的對策刻不容緩。家長在把孩子交給「文字拼湊機器」之前，應三思而後行，不論那台機器是否縫在泰迪熊裡。聊天機器人應設有並確實執行年齡限制；政府不該再給 AI 公司如同當年社群平台般的寬鬆待遇。教師若仍相信在家完成的作文可信，無疑是自欺欺人；在 AI 時代，更多校內評量勢在必行。 更長遠的挑戰，是如何保留 AI 可能從童年生活中抹去的社會化過程，而學校正是最佳場域。學校應在證實有效的情況下，善用個人化教學；同時，也必須加倍努力教授機器無法取代的能力：辯論、分歧、合作，甚至欣賞那些不像聊天機器人般阿諛奉承的人。 快樂的王子，空洞的王國 學校也應強化作為探索中心的角色。若 AI 不斷給孩子他們想要的內容，那麼學校更應提供機會，讓他們接觸經驗之外的人與思想。演算法式個人化，若引導人們始終停留在起點所在的「車道」，將成為社會流動的強大障礙；若貧困學校僅以聊天機器人作為廉價的人類教師替代品，不平等恐將擴大。 AI 無疑具備改善教育、豐富娛樂的潛力，或許終有一天，能讓每個孩子都過上如王室般的生活。但真正的特權，可能屬於那些知道何時該把它關掉的家長與教師。 ■ --Hosting provided by SoundOn