M\u1ea5y h\u00f4m tr\u01b0\u1edbc con g\u00e1i h\u1ecfi t\u00f4i \u201cM\u1eb9 c\u00f3 bi\u1ebft m\u1ea1ng x\u00e3 h\u1ed9i t\u1eebng hot c\u00e1i trend Cake or real kh\u00f4ng?\u201d. T\u00f4i c\u00f3 nh\u1edb, \u0111\u00f3 l\u00e0 khi \u0111\u00f4i gi\u00e0y, c\u1ee5c g\u1ea1ch, t\u00fai x\u00e1ch t\u01b0\u1edfng th\u1eadt m\u00e0 c\u1eaft ra ho\u00e1 ra l\u1ea1i l\u00e0 b\u00e1nh. Ch\u00fang ta xem v\u00e0 b\u1eadt c\u01b0\u1eddi.<\/p>\n\n\n\n
Nh\u01b0ng con g\u00e1i t\u00f4i l\u1ea1i b\u1ea3o trend \u0111\u00f3 \u0111\u00e3 c\u0169 r\u1ed3i – b\u00e2y gi\u1edd l\u00e0 \u201cAI or real?\u201d. T\u1eeb c\u00e1 s\u1ea5u l\u00ean b\u1edd, r\u1eafn \u1edf tr\u00ean gi\u01b0\u1eddng, m\u00e8o n\u1ea5u \u0103n, nh\u1eefng s\u1ea3n ph\u1ea9m qu\u1ea3ng c\u00e1o \u0111\u1eb9p ho\u00e0n h\u1ea3o.v.v m\u1ecdi video do AI gi\u1ea3 l\u1eadp tr\u00e0n lan. V\u00e0 ch\u1eafc ch\u1eafn d\u00f9 ta c\u00f3 c\u1ed1 soi \u0111\u1ebfn m\u1ea5y c\u0169ng r\u1ea5t kh\u00f3 \u0111\u1ec3 nh\u1eadn bi\u1ebft \u0111\u00e2u l\u00e0 gi\u1ea3 – th\u1eadt. Gi\u1edd th\u00ec ch\u00fang ta xem c\u00e1c n\u1ed9i dung do AI gi\u1ea3 l\u1eadp v\u00e0 kh\u00f3 m\u00e0 c\u01b0\u1eddi. S\u00e2u trong ta l\u00e0 s\u1ef1 lo s\u1ee3 v\u00e0 ho\u00e0i nghi v\u1ec1 t\u1ea5t th\u1ea3y. Ta nh\u1eadn ra: trong k\u1ef7 nguy\u00ean AI, th\u1ee9 khan hi\u1ebfm nh\u1ea5t ch\u00ednh l\u00e0 S\u1ef0 TH\u1eacT!<\/p>\n\n\n\n
AI \u0111\u00e3 bi\u1ebfn gi\u1ea3 th\u00e0nh \u201ch\u00e0ng c\u00f4ng nghi\u1ec7p\u201d: review gi\u1ea3, chatbot gi\u1ea3 m\u1ea1o, deepfake tung ho\u00e0nh. Chi ph\u00ed cho nh\u1eefng ngu\u1ed3n gi\u1ea3 n\u00e0y g\u1ea7n nh\u01b0 b\u1eb1ng 0 khi\u1ebfn n\u00f3 c\u00e0ng sinh s\u00f4i nh\u01b0 c\u1ecf d\u1ea1i. Ch\u00fang ta – trong vai th\u00e0nh vi\u00ean c\u1ee7a m\u1ed9t nh\u00f3m x\u00e3 h\u1ed9i n\u00e0o \u0111\u00f3, l\u00e0 ng\u01b0\u1eddi ti\u00eau d\u00f9ng c\u1ee7a nh\u1eefng th\u01b0\u01a1ng hi\u1ec7u n\u00e0o \u0111\u00f3: m\u1ec7t m\u1ecfi, nghi ng\u1edd, m\u1ea5t ki\u00ean nh\u1eabn.<\/p>\n\n\n\n
V\u00e0 t\u00e2m l\u00fd h\u1ecdc t\u00e2m l\u00fd x\u00e3 h\u1ed9i g\u1ecdi \u0111\u00e2y l\u00e0 hi\u1ec7u \u1ee9ng khan hi\u1ebfm(Scarcity effect): khi m\u1ed9t \u0111i\u1ec1u tr\u1edf n\u00ean hi\u1ebfm, gi\u00e1 tr\u1ecb c\u1ea3m nh\u1eadn v\u00e0 \u0111\u1ed9ng l\u1ef1c t\u00ecm ki\u1ebfm c\u1ee7a ta t\u0103ng l\u00ean. Khi gi\u1ea3 lan tr\u00e0n, s\u1ef1 th\u1eadt b\u1ed7ng th\u00e0nh \u201cxa x\u1ec9 ph\u1ea9m\u201d. V\u00e0 ch\u00ednh v\u00ec hi\u1ebfm, n\u00f3 \u0111\u01b0\u1ee3c \u0111\u1ecbnh gi\u00e1 cao h\u01a1n bao gi\u1edd h\u1ebft.<\/p>\n\n\n\n
M\u1eb7t kh\u00e1c, m\u1ed7i ch\u00fang ta c\u00f2n r\u01a1i v\u00e0o m\u1ed9t hi\u1ec7n t\u01b0\u1ee3ng g\u00e1nh n\u1eb7ng t\u00e2m l\u00fd kh\u00e1c: M\u1ec7t m\u1ecfi khi ra quy\u1ebft \u0111\u1ecbnh (decision fatigue). M\u1ed7i ng\u00e0y ta ph\u1ea3i ph\u00e2n bi\u1ec7t th\u1eadt gi\u1ea3, khi\u1ebfn n\u00e3o ki\u1ec7t s\u1ee9c. Khi m\u1ec7t m\u1ecfi, con ng\u01b0\u1eddi t\u00ecm nh\u1eefng n\u01a1i an to\u00e0n \u0111\u1ec3 \u201c\u0111\u1ee1 ph\u1ea3i ngh\u0129 nhi\u1ec1u\u201d. Khi n\u00e0y, nh\u1eefng th\u01b0\u01a1ng hi\u1ec7u v\u00e0 n\u1ec1n t\u1ea3ng \u01b0u ti\u00ean n\u1ed9i dung do ng\u01b0\u1eddi th\u1eadt t\u1ea1o v\u00e0 c\u00f3 th\u1ec3 ki\u1ec3m ch\u1ee9ng s\u1ef1 th\u1eadt, nh\u01b0 c\u00e2u chuy\u1ec7n c\u1ee7a Locket – t\u01b0\u1edfng ch\u1eebng nh\u01b0 l\u00e0 m\u1ed9t \u1ee9ng d\u1ee5ng qu\u00e1 \u0111\u1ed7i \u0111\u01a1n gi\u1ea3n quay l\u1ea1i c\u00e1ch \u0111\u00e2y 20 n\u0103m nh\u01b0ng Locket t\u1ea1o c\u1ea3m gi\u00e1c an to\u00e0n nh\u1eadn th\u1ee9c \u1edf th\u1eddi \u0111\u1ea1i n\u00e0y! N\u00f3 b\u1eaft \u0111\u1ea7u cho m\u1ed9t v\u0103n h\u00f3a \u201cauthenticity\u201d c\u1ee7a ng\u01b0\u1eddi tr\u1ebb (Gen Z, Alpha) \u0111\u1ec1 cao t\u00ednh ch\u00e2n th\u1eadt: raw & real.<\/p>\n\n\n\n
B\u1ea1n bi\u1ebft \u0111\u1ea5y, x\u00e3 h\u1ed9i lu\u00f4n v\u1eadn h\u00e0nh d\u1ef1a tr\u00ean ni\u1ec1m tin (ti\u1ec1n, lu\u1eadt, th\u01b0\u01a1ng hi\u1ec7u \u0111\u1ec1u l\u00e0 c\u00e1c c\u1ea5u tr\u00fac ni\u1ec1m tin). Khi c\u00e1i gi\u1ea3 lan tr\u00e0n, ni\u1ec1m tin t\u1eadp th\u1ec3 lung lay, k\u00e9o theo r\u1ee7i ro x\u00e3 h\u1ed9i. S\u1ef1 th\u1eadt kh\u00f4ng ch\u1ec9 l\u00e0 th\u00f4ng tin, m\u00e0 l\u00e0 ni\u1ec1m tin t\u1eadp th\u1ec3. Trong th\u1ebf gi\u1edbi h\u00f4m nay, s\u1ef1 th\u1eadt kh\u00f4ng ch\u1ec9 l\u00e0 gi\u00e1 tr\u1ecb, n\u00f3 c\u00f2n l\u00e0 ti\u1ec1n t\u1ec7.<\/p>\n\n\n\n
Nghe ngh\u1ecbch l\u00fd, nh\u01b0ng ch\u00ednh nh\u1eefng n\u01a1i \u0111\u1ec3 C\u00e1i gi\u1ea3 lan nhanh nh\u01b0 d\u1ecbch b\u1ec7nh – l\u1ea1i l\u00e0 nh\u1eefng \u201cng\u01b0\u1eddi b\u00e1n s\u1ef1 th\u1eadt\u201d l\u1edbn nh\u1ea5t. V\u00ec sao l\u1ea1i th\u1ebf? V\u00ec con ng\u01b0\u1eddi ng\u00e0y c\u00e0ng m\u1ec7t m\u1ecfi v\u1edbi nh\u1eefng th\u00f4ng tin gi\u1ea3 m\u1ea1o – h\u1ecd s\u1ebd r\u1eddi b\u1ecf nh\u1eefng n\u1ec1n t\u1ea3ng hay th\u01b0\u01a1ng hi\u1ec7u qu\u00e1 nhi\u1ec1u th\u00f4ng tin gi\u1ea3 n\u00e0y. C\u00e1c n\u1ec1n t\u1ea3ng x\u00e3 h\u1ed9i, th\u01b0\u01a1ng hi\u1ec7u c\u00e0ng ph\u1ea3i ch\u1ea1y \u0111ua \u0111\u1ec3 gi\u1eef t\u00ednh ch\u00e2n th\u1eadt c\u1ee7a n\u1ed9i dung kh\u00f4ng ch\u1ec9 v\u00ec \u0111\u1ea1o \u0111\u1ee9c, m\u00e0 v\u00ec ni\u1ec1m tin = doanh thu. N\u1ebfu to\u00e0n n\u1ed9i dung gi\u1ea3, ng\u01b0\u1eddi d\u00f9ng, kh\u00e1ch h\u00e0ng s\u1ebd r\u1eddi \u0111i. C\u00f9ng nh\u00ecn c\u00e1ch m\u00e0 Google, Tiktok hay Open Ai v\u1eeba lan truy\u1ec1n n\u1ed9i dung fake, v\u1eeba \u201cb\u00e1n s\u1ef1 th\u1eadt\u201d:<\/p>\n\n\n\n
-Google t\u1eebng l\u00e0 n\u01a1i m\u00e0 \u0111\u1ec3 l\u1ecdt n\u1ed9i dung gi\u1ea3 kh\u00f4ng ki\u1ec3m ch\u1ee9ng lan tr\u00e0n. Nh\u01b0ng hi\u1ec7n nay Google \u0111ang \u0111i\u1ec1u ch\u1ec9nh h\u1ec7 th\u1ed1ng \u0111\u1ec3 \u01b0u ti\u00ean n\u1ed9i dung h\u1eefu \u00edch v\u00e0 \u0111\u00e1ng tin; E-E-A-T (Experience – Expertise – Authoritativeness – Trustworthiness) l\u00e0 khung \u0111\u00e1nh gi\u00e1 gi\u00fap h\u1ecd ki\u1ec3m so\u00e1t ch\u1ea5t l\u01b0\u1ee3ng. Ngh\u0129a l\u00e0 khi b\u1ea1n t\u00ecm ki\u1ebfm m\u1ed9t t\u1eeb kho\u00e1 tr\u00ean google, ch\u1ec9 c\u00e1c t\u1eeb kho\u00e1 t\u1eeb c\u00e1c trang web c\u00f3 E-E-A-T cao m\u1edbi \u0111\u01b0\u1ee3c \u01b0u ti\u00ean hi\u1ec3n th\u1ecb. C\u00e1c n\u1ec1n t\u1ea3ng, th\u01b0\u01a1ng hi\u1ec7u mu\u1ed1n \u0111\u01b0\u1ee3c hi\u1ec3n th\u1ecb bu\u1ed9c ph\u1ea3i t\u1ea1o n\u1ed9i dung th\u1ef1c t\u1ebf (test s\u1ea3n ph\u1ea9m, chia s\u1ebb case study, review th\u1eadt), thay v\u00ec ch\u1ec9 vi\u1ebft n\u1ed9i dung gi\u1ea3 t\u1ea1o, k\u00e9m ch\u1ea5t l\u01b0\u1ee3ng.<\/p>\n\n\n\n
-Hay TikTok, m\u1ed9t n\u1ec1n t\u1ea3ng v\u1edbi c\u00e1c video ng\u1eafn v\u00e0 gi\u1ea3 m\u1ea1o tr\u00e0n lan \u0111\u00e3 si\u1ebft thu\u1eadt to\u00e1n \u0111\u1ec3 \u01b0u ti\u00ean n\u1ed9i dung ch\u00e2n th\u1ef1c, ng\u01b0\u1eddi th\u1eadt, quay b\u1eb1ng \u0111i\u1ec7n tho\u1ea1i, h\u1ea1n ch\u1ebf n\u1ed9i dung AI t\u1ea1o.<\/p>\n\n\n
-OpenAI ch\u00ednh l\u00e0 n\u1ec1n t\u1ea3ng tung ra ChatGPT v\u00e0 c\u00e1c \u1ee9ng d\u1ee5ng t\u1ea1o video, h\u00ecnh \u1ea3nh b\u1eb1ng AI – l\u00e0 c\u00e1i n\u00f4i \u0111\u1ec3 ng\u01b0\u1eddi d\u00f9ng sinh ra nh\u1eefng n\u1ed9i dung gi\u1ea3 tr\u00e0n lan – l\u1ea1i v\u1eeba ph\u00e1t tri\u1ec3n h\u1ec7 th\u1ed1ng watermark nh\u01b0 Content Credentials (si\u00eau d\u1eef li\u1ec7u ngu\u1ed3n g\u1ed1c) nh\u1eadn bi\u1ebft n\u1ed9i dung \u0111\u01b0\u1ee3c t\u1ea1o ra b\u1eb1ng AI.<\/p>\n\n\n
C\u00f3 th\u1ec3 n\u00f3i khi AI ph\u00e1t tri\u1ec3n, c\u00e1c c\u00f4ng ty Big Tech (Google, TikTok, OpenAI\u2026) ch\u00ednh l\u00e0 ngu\u1ed3n tung ra l\u00e0n s\u00f3ng n\u1ed9i dung gi\u1ea3 deepfake, n\u1ed9i dung AI spam, review gi\u1ea3 m\u1ea1o h\u00e0ng lo\u1ea1t. Hi\u1ec7n nay c\u00e1c Big tech n\u00e0y l\u1ea1i ch\u00ednh l\u00e0 nh\u1eefng n\u01a1i \u0111ang ch\u1ea1y \u0111ua h\u1ebft s\u1ee9c \u0111\u1ec3 ki\u1ec3m \u0111\u1ecbnh s\u1ef1 th\u1eadt, ti\u00ean phong tung ra c\u00e1c c\u00f4ng c\u1ee5 AI detector, watermark, blockchain x\u00e1c minh, c\u00f4ng c\u1ee5 ki\u1ec3m ch\u1ee9ng s\u1ef1 th\u1eadt. C\u00e2u chuy\u1ec7n th\u1eadt quen thu\u1ed9c, h\u01a1i gi\u1ed1ng tung ra d\u1ecbch b\u1ec7nh r\u1ed3i b\u00e1n vacxin ph\u1ea3i kh\u00f4ng?<\/p>\n\n\n\n
V\u00ec kh\u00f4ng ch\u1ec9 v\u00ec doanh thu, kh\u00e1ch h\u00e0ng hay \u0111\u1ea1o \u0111\u1ee9c m\u00e0 h\u1ecd tung ra c\u00e1c c\u00f4ng c\u1ee5 ki\u1ec3m ch\u1ee9ng s\u1ef1 th\u1eadt. C\u00f2n c\u00f3 m\u1ed9t l\u00fd do l\u1edbn h\u01a1n th\u1ebf! \u0110\u00f3 l\u00e0 ch\u00ednh tr\u1ecb.<\/p>\n\n\n\n
T\u1eeb th\u1eddi c\u1ed5 \u0111\u1ea1i, c\u00e1c nh\u00e0 c\u1ea7m quy\u1ec1n n\u1eafm quy\u1ec1n v\u00ec h\u1ecd c\u00f3 kh\u1ea3 n\u0103ng ki\u1ec3m so\u00e1t \u201cc\u00e2u chuy\u1ec7n ch\u00ednh th\u1ed1ng\u201d (v\u00ed d\u1ee5: s\u1eed s\u00e1ch, t\u00f4n gi\u00e1o, lu\u1eadt l\u1ec7). Ngh\u0129a l\u00e0: ai c\u00f3 quy\u1ec1n \u0111\u1ecbnh ngh\u0129a \u201cs\u1ef1 th\u1eadt\u201d th\u00ec ng\u01b0\u1eddi \u0111\u00f3 c\u00f3 quy\u1ec1n l\u1ef1c ch\u00ednh tr\u1ecb. T\u01b0\u01a1ng t\u1ef1 nh\u01b0 b\u00e1o ch\u00ed t\u1eebng b\u1ecb ki\u1ec3m so\u00e1t ch\u1eb7t ch\u1ebd ch\u00ednh ph\u1ee7\/nh\u00f3m quy\u1ec1n l\u1ef1c \u0111\u1ecbnh ngh\u0129a s\u1ef1 th\u1eadt. C\u00e1c c\u00f4ng ty c\u00f4ng ngh\u1ec7 Big Tech n\u1eafm d\u1eef li\u1ec7u c\u00e1 nh\u00e2n th\u1ebf k\u1ef7 21. V\u00e0 gi\u1edd l\u00e0 cu\u1ed9c \u0111ua n\u1eafm gi\u1eef s\u1ef1 th\u1eadt s\u1ed1 th\u1ebf k\u1ef7 22.<\/p>\n\n\n\n
V\u1ec1 t\u00e2m l\u00fd, ch\u00fang ta m\u1eb7c \u0111\u1ecbnh tin (truth-default) v\u00e0 d\u1ec5 \u0111i theo d\u1ea5u v\u1ebft \u0111\u00e1m \u0111\u00f4ng (social proof). Trong th\u1eddi \u0111\u1ea1i AI, th\u00f4ng tin nhi\u1ec5u lo\u1ea1n (Truth Decay) ng\u01b0\u1eddi d\u00e2n kh\u00f4ng c\u00f2n bi\u1ebft tin \u0111\u00e2u l\u00e0 s\u1ef1 th\u1eadt, d\u1ec5 t\u1ea1o ra kh\u1ee7ng ho\u1ea3ng ni\u1ec1m tin, v\u00ec th\u1ebf x\u00e3 h\u1ed9i t\u1ef1 nhi\u00ean \u0111\u00f2i h\u1ecfi m\u1ed9t \u201ctr\u1ecdng t\u00e0i\u201d \u0111\u1ec3 ch\u1ee9ng th\u1ef1c \u0111\u00e2u l\u00e0 s\u1ef1 th\u1eadt. L\u00fac \u0111\u00f3 s\u1ebd nguy hi\u1ec3m khi \u201cs\u1ef1 th\u1eadt\u201d c\u00f3 th\u1ec3 b\u1ecb thao t\u00fang v\u00ec l\u1ee3i \u00edch ch\u00ednh tr\u1ecb\/kinh t\u1ebf c\u1ee7a \u201cTr\u1ecdng t\u00e0i\u201d. \u201cTr\u1ecdng t\u00e0i\u201d thao t\u00fang \u0111\u00e2u l\u00e0 s\u1ef1 th\u1eadt, \u0111\u00e2u l\u00e0 gi\u1ea3. V\u00ed d\u1ee5: n\u1ebfu \u201ctr\u1ecdng t\u00e0i\u201d l\u00e0 m\u1ed9t ch\u00ednh ph\u1ee7 th\u00ec h\u1ecd c\u00f3 th\u1ec3 g\u1eafn nh\u00e3n \u201cfake news\u201d cho c\u00e1c th\u00f4ng tin ch\u1ed1ng \u0111\u1ed1i, bi\u1ebfn s\u1ef1 th\u1eadt th\u00e0nh v\u0169 kh\u00ed ch\u00ednh tr\u1ecb.<\/p>\n\n\n\n
V\u00ec v\u1eady hi\u1ec7n nay, quy\u1ec1n \u0111\u1ecbnh ngh\u0129a \u201c\u0111\u00e2u l\u00e0 s\u1ef1 th\u1eadt\u201d n\u1eb1m trong tay ba th\u1ebf l\u1ef1c: Nh\u00e0 n\u01b0\u1edbc \u0111\u1eb7t lu\u1eadt; N\u1ec1n t\u1ea3ng l\u00e0m “ng\u01b0\u1eddi g\u00e1c c\u1ed5ng”; v\u00e0 C\u1ed9ng \u0111\u1ed3ng l\u00e0 l\u1edbp \u0111\u1ed3ng ki\u1ec3m. Nh\u00ecn theo t\u00e2m l\u00fd h\u1ecdc ch\u00ednh tr\u1ecb, tranh ch\u1ea5p n\u00e0y xoay quanh t\u00ednh ch\u00ednh \u0111\u00e1ng c\u1ee7a ngu\u1ed3n: ai \u0111\u01b0\u1ee3c c\u00f4ng nh\u1eadn l\u00e0 “tr\u1ecdng t\u00e0i” s\u1ebd \u0111\u1eb7t chu\u1ea9n \u0111\u00fang – sai cho m\u1ecdi th\u00f4ng tin:<\/p>\n\n\n\n
1 – C\u00e1c c\u00f4ng ty c\u00f4ng ngh\u1ec7 l\u1edbn – Big Tech nh\u01b0 Google, TikTok, OpenAI v\u1eeba l\u00e0 n\u01a1i s\u1ea3n sinh, nu\u00f4i d\u01b0\u1ee1ng gi\u1ea3 m\u1ea1o b\u00f9ng ph\u00e1t, v\u1eeba kh\u00f4ng ng\u1eebng t\u1ea1o c\u00e1c c\u00f4ng c\u1ee5 gi\u1ea3i ph\u00e1p ch\u1ee9ng th\u1ef1c n\u1ed9i dung s\u1ed1 \u0111\u01b0\u1ee3c t\u1ea1o b\u1eb1ng AI \u0111\u1ec3 ki\u1ec3m so\u00e1t \u201cs\u1ef1 th\u1eadt\u201d. Ngo\u00e0i ra li\u00ean minh nhi\u1ec1u h\u00e3ng c\u00f4ng ngh\u1ec7 l\u1edbn \u0111\u00e3 x\u00e2y d\u1ef1ng Ban \u0111i\u1ec1u h\u00e0nh (Open AI, Adobe, Intel, Microsoft, Truepic.v.v) cho ra \u0111\u1eddi chu\u1ea9n C2PA ( Coalition for Content Provenance and Authenticity) – chu\u1ea9n k\u1ef9 thu\u1eadt m\u1edf gi\u00fap g\u1eafn \u201cl\u00fd l\u1ecbch s\u1ed1\u201d v\u00e0o \u1ea3nh, video, \u00e2m thanh\u2026 \u0111\u1ec3 cho bi\u1ebft n\u1ed9i dung \u0111\u1ebfn t\u1eeb \u0111\u00e2u, do ai t\u1ea1o\/ch\u1ec9nh s\u1eeda, b\u1eb1ng c\u00f4ng c\u1ee5 n\u00e0o, v\u00e0 cho ph\u00e9p m\u1ecdi ng\u01b0\u1eddi x\u00e1c minh t\u00ednh to\u00e0n v\u1eb9n c\u1ee7a th\u00f4ng tin \u0111\u00f3.<\/p>\n\n\n\n
2 – Ch\u00ednh ph\u1ee7: \u1ede ch\u00e2u \u00c2u, AI Act \u0111\u00e3 c\u00f3 hi\u1ec7u l\u1ef1c – trong \u0111\u00f3 \u0110i\u1ec1u 50(4) y\u00eau c\u1ea7u g\u1eafn nh\u00e3n n\u1ed9i dung deepfake; c\u00e1c m\u1ed1c \u00e1p d\u1ee5ng \u0111ang cu\u1ed1n chi\u1ebfu \u0111\u1ebfn 2026 – 2027. \u0110\u00e2y l\u00e0 n\u1ed7 l\u1ef1c ch\u00ednh tr\u1ecb nh\u1eb1m ki\u1ec3m so\u00e1t \u201cs\u1ef1 th\u1eadt s\u1ed1\u201d. EU mu\u1ed1n bi\u1ebfn \u201cki\u1ec3m so\u00e1t s\u1ef1 th\u1eadt s\u1ed1\u201d th\u00e0nh m\u1ed9t ngh\u0129a v\u1ee5 ph\u00e1p l\u00fd. Thay v\u00ec \u0111\u1ec3 Google, TikTok t\u1ef1 quy\u1ebft \u0111\u1ecbnh, ch\u00ednh ph\u1ee7 (v\u00e0 lu\u1eadt ph\u00e1p) s\u1ebd \u0111\u00f3ng vai \u201ctr\u1ecdng t\u00e0i s\u1ef1 th\u1eadt\u201d. Ch\u00ednh tr\u1ecb h\u00f3a \u1edf \u0111\u00e2y: Nh\u00e0 n\u01b0\u1edbc b\u01b0\u1edbc v\u00e0o ki\u1ec3m so\u00e1t, kh\u00f4ng c\u00f2n \u0111\u1ec3 Big Tech t\u1ef1 tung t\u1ef1 t\u00e1c.<\/p>\n\n\n\n
3 – C\u1ed9ng \u0111\u1ed3ng – ch\u00ednh l\u00e0 ch\u00fang ta: t\u1eeb n\u0103m 2021- 2023 Elon Musk ra m\u1eaft ch\u1ee9c n\u0103ng Community Notes \u0111\u1ec3 c\u1ed9ng \u0111\u1ed3ng x\u00e1c minh c\u00e1c n\u1ed9i dung th\u1eadt c\u00f3 h\u1eefu \u00edch tr\u00ean n\u1ec1n t\u1ea3ng X, t\u01b0\u01a1ng t\u1ef1 ch\u1ee9c n\u0103ng n\u00e0y, Tiktok th\u1eed nghi\u1ec7m Footnotes t\u1eeb 4\/2025 v\u00e0 m\u1edf r\u1ed9ng t\u1ea1i M\u1ef9 7\/2025. \u0110\u00e2y l\u00e0 m\u1ed9t d\u1ea1ng \u201cch\u00ednh tr\u1ecb ho\u00e1 s\u1ef1 th\u1eadt\u201d v\u00ec trao quy\u1ec1n cho c\u1ed9ng \u0111\u1ed3ng thay v\u00ec ch\u00ednh ph\u1ee7 hay t\u1eadp \u0111o\u00e0n. C\u00e1c Big tech c\u00f3 l\u1ebd c\u1ea3m th\u1ea5y vi\u1ec7c \u0111\u1ed9c chi\u1ebfm quy\u1ec1n ki\u1ec3m \u0111\u1ecbnh s\u1ef1 th\u1eadt l\u00e0 kh\u00f3 kh\u0103n khi gi\u1eb1ng co v\u1edbi ch\u00ednh ph\u1ee7, v\u00ec v\u1eady h\u1ecd \u0111ang th\u1eed bi\u1ebfn \u201cki\u1ec3m ch\u1ee9ng s\u1ef1 th\u1eadt\u201d th\u00e0nh quy\u1ec1n c\u1ee7a c\u1ed9ng \u0111\u1ed3ng thay v\u00ec giao cho ch\u00ednh ph\u1ee7. \u0110\u00e2y l\u00e0 m\u1ed9t tuy\u00ean ng\u00f4n: \u201cAi c\u00f3 quy\u1ec1n \u0111\u1ecbnh ngh\u0129a s\u1ef1 th\u1eadt? Kh\u00f4ng ph\u1ea3i ch\u00ednh ph\u1ee7, kh\u00f4ng ph\u1ea3i Big Tech, m\u00e0 l\u00e0 c\u1ed9ng \u0111\u1ed3ng.\u201d \u0110i\u1ec1u n\u00e0y tr\u00e1i ng\u01b0\u1ee3c v\u1edbi EU, n\u01a1i m\u00e0 ch\u00ednh ph\u1ee7 mu\u1ed1n ki\u1ec3m so\u00e1t s\u1ef1 th\u1eadt s\u1ed1.<\/p>\n\n\n\n
N\u00f3i n\u00f4m na: S\u1ef1 th\u1eadt ch\u00ednh l\u00e0 m\u1ed9t \u201ct\u00e0i nguy\u00ean ch\u00ednh tr\u1ecb\u201d m\u1edbi. Ai gi\u1eef ch\u00eca kho\u00e1 s\u1ef1 th\u1eadt, ng\u01b0\u1eddi \u0111\u00f3 c\u00f3 th\u1ec3 \u0111\u1ecbnh h\u00ecnh nh\u1eadn th\u1ee9c x\u00e3 h\u1ed9i, \u1ea3nh h\u01b0\u1edfng c\u1ea3 b\u1ea7u c\u1eed l\u1eabn th\u1ecb tr\u01b0\u1eddng. Th\u1eadt nguy hi\u1ec3m n\u1ebfu s\u1ef1 th\u1eadt b\u1ecb ph\u1ee5 thu\u1ed9c v\u00e0o 1 \u201c tr\u1ecdng t\u00e0i\u201d, v\u1eady th\u00ec \u201ctr\u1ecdng t\u00e0i\u201d \u0111\u00f3 s\u1ebd gi\u1ed1ng nh\u01b0 \u00f4ng ch\u1ee7 duy nh\u1ea5t s\u1edf h\u1eefu b\u1ed9 l\u1ecdc n\u01b0\u1edbc s\u1ea1ch, c\u00f2n m\u1ecdi ng\u01b0\u1eddi ph\u1ea3i tr\u1ea3 ti\u1ec1n u\u1ed1ng.<\/p>\n\n\n\n
C\u1ee9 theo \u0111\u00e0 n\u00e0y, m\u1ed9t c\u00e2u h\u1ecfi \u0111\u1eb7t ra l\u00e0 s\u1ef1 th\u1eadt tuy\u1ec7t \u0111\u1ed1i c\u00f2n t\u1ed3n t\u1ea1i trong \u0111\u1eddi s\u1ed1ng s\u1ed1, hay ch\u1ec9 c\u00f2n s\u1ef1 th\u1eadt t\u01b0\u01a1ng \u0111\u1ed1i do s\u1ed1 \u0111\u00f4ng \u0111\u1ed3ng thu\u1eadn? V\u00e0 s\u1ef1 th\u1eadt m\u00e0 \u0111\u00e1m \u0111\u00f4ng \u0111\u1ed3ng thu\u1eadn c\u00f3 ph\u1ea3i do ch\u00fang ta ch\u1ee7 \u0111\u1ed9ng hay v\u1eabn do Big tech trao quy\u1ec1n?<\/p>\n\n\n\n
C\u00f3 l\u1ebd, tr\u01b0\u1edbc khi h\u1ecfi \u201cl\u00e0m sao \u0111\u1ec3 bi\u1ebft \u0111\u00e2u l\u00e0 s\u1ef1 th\u1eadt?\u201d, ta n\u00ean h\u1ecfi: t\u1ea1i sao n\u00f3 l\u1ea1i khan hi\u1ebfm \u0111\u1ebfn th\u1ebf? C\u00f3 ph\u1ea3i v\u00ec th\u1ebf gi\u1edbi AI \u0111ang ti\u1ebfp tay cho nh\u1eefng th\u00f4ng tin gi\u1ea3 m\u1ea1o nhi\u1ec1u h\u01a1n? V\u1eady th\u00ec AI ti\u1ebfp tay cho ai? Ch\u00fang l\u00e0 ch\u00fang ta – nh\u1eefng con ng\u01b0\u1eddi s\u1eed d\u1ee5ng ch\u00fang. Ch\u00ednh con ng\u01b0\u1eddi \u0111ang l\u00e0m cho AI tr\u1edf th\u00e0nh c\u00f4ng c\u1ee5 \u0111\u1ec3 t\u1ea1o ra c\u00e1c n\u1ed9i dung gi\u1ea3, c\u00f4ng c\u1ee5 l\u1eeba \u0111\u1ea3o nhau, \u0111\u1ec3 thu l\u1ee3i m\u1ed9t c\u00e1ch sai tr\u00e1i.<\/p>\n\n\n\n
S\u1ef1 th\u1eadt s\u1ebd th\u00f4i kh\u00f4ng c\u00f2n khan hi\u1ebfm khi ta cho n\u00f3 \u0111\u1ea5t s\u1ed1ng, khi ta c\u00f3 th\u00eam k\u1ef7 lu\u1eadt v\u00e0 l\u00f2ng can \u0111\u1ea3m: d\u00e1m ch\u1eadm l\u1ea1i, s\u1ed1ng th\u1eadt v\u1edbi kh\u1ea3 n\u0103ng, t\u00e0i nguy\u00ean c\u1ee7a ch\u00ednh m\u00ecnh, d\u00e1m nghi ng\u1edd c\u1ea3 \u0111i\u1ec1u m\u00ecnh mu\u1ed1n tin, v\u00e0 d\u00e1m nh\u1eadn sai. \u0110\u1eebng \u0111\u1ee3i \u0111\u1ebfn l\u00fac c\u00e1c n\u1ec1n t\u1ea3ng c\u00f4ng ngh\u1ec7 s\u00e0ng l\u1ecdc c\u00e1c n\u1ed9i dung fake c\u1ee7a ch\u00fang ta t\u1ea1o ra. Tr\u01b0\u1edbc khi t\u1ea1o n\u1ed9i dung hay b\u1ea5m chia s\u1ebb b\u1ea5t k\u1ef3 \u0111i\u1ec1u g\u00ec h\u00e3y ch\u1ee7 \u0111\u1ed9ng c\u00f3 tr\u00e1ch nhi\u1ec7m v\u00e0 t\u1ef1 h\u1ecfi v\u1ec1 ngu\u1ed3n g\u1ed1c c\u1ee7a n\u1ed9i dung \u0111\u00f3. H\u00e3y ch\u1ec9 s\u1eed d\u1ee5ng c\u00e1c c\u00f4ng c\u1ee5 AI \u0111\u00fang c\u00e1ch v\u00e0o vi\u1ec7c n\u00e2ng cao tri th\u1ee9c cho con ng\u01b0\u1eddi.<\/p>\n\n\n\n
<\/p>\n\n\n\n