AI & C ea i i y?
Why AI canno c ea e A : Chiang's a gumen s e isi ed
Science ic ion w i e Ted Chiang has a gued why AI canno c ea e a (Chiang, 2024). I ecommend
eading his a icle (and o cou se books!). I ha e p oposed some coun e a gumen s (He nández-
Fe nández, 2025), which, like his ex , a e mean o p o oke hough . Chiang’s a gumen s agains AI’s
a is ic capabili ies can be summa ized as ollows (Chiang, 2024; He nández-Fe nández, 2025):
1. The Mic odecision A gumen – A in ol es nume ous conscious and unconscious decisions.
Chiang a gues ha AI elies solely on s a is ical models and imi a ion, p oducing p edic able and
medioc e esul s, aligning wi h Wagensbe g’s (2017) no ion ha he opposi e o c ea i i y is
medioc i y.
2. The Lack o In en ionali y A gumen – T ue a equi es in en ion. Since AI has no
emo ions, in ellec ual joy, o desi es, i canno c ea e wi h meaning ul in en .
3. The E o and P ocess A gumen – The c ea i e p ocess and e o a e in eg al o a is ic
alue. AI gene a es con en e o lessly, which Chiang belie es diminishes a is ic wo h.
4. The Gene al In elligence and E iciency A gumen – AI s uggles o acqui e new skills
e icien ly compa ed o humans. Despi e excelling a games like Go, i equi es as da ase s,
whe eas humans lea n mo e in ui i ely—key o a is ic c ea ion.
5. The Lack o O iginali y A gumen – AI is gene a i e bu no c ea i e. I s ou pu s, buil on
au oma ed mechanisms, lack ue o iginali y, esul ing in s a is ically a e age wo k.
These poin s lead o discussions in class ooms and acul y mee ings abou a (and a is ic educa ion),
aising coun e a gumen s ha may con ibu e o u u e esea ch and deba e (He nández-Fe nández,
2025):
a. AI and A is ic Mic odecisions – While AI's decisions a e based on p io da a and algo i hms
(wi h inhe en biases), u u e ad ancemen s may allow AI o be e eplica e human a is ic p ocesses.
This aises ques ions abou hyb id human-AI c ea i e in e ac ions (mic odecisions in p omp ing, in AI-
human eedback, e c.).
b. In en ionali y and E hical Use o AI – While AI lacks in en , educa o s and a is s can use i
exp essi ely h ough echniques like p omp ing (He nández-Fe nández & Fe e -i-Cancho, 2023).
Recognizing he link be ween echnology and i s usage is c ucial o a oiding misconcep ions o AI
au onomy (Diéguez, 2024). Teache s play a key ole in guiding s uden s owa d esponsible and
in en ional c ea i i y.
c. Op imizing C ea i e E o – AI can handle epe i i e echnical asks, allowing a is s o ocus on
concep ual and c ea i e aspec s, much like how pho og aphy e olu ionized cinema. AI may
democ a ize digi al c ea ion, enabling use s o gene a e so wa e wi hou coding knowledge o p oduce
g aphic wo ks o 3D-p in ed pieces (He nández-Fe nández, 2023).
d. Sus ainabili y and AI E iciency in A – While AI cu en ly has a high en i onmen al cos
(C aw o d, 2021), u u e sus ainable de elopmen s may enable esponsible in eg a ion in o a
educa ion, os e ing new c ea i e p ocesses.
e. Human-Machine Collabo a i e C ea i i y – AI can ex end he bounda ies o collabo a i e a is ic
c ea ion. In educa ion, AI could se e as a c ea i e pa ne a he han me ely a ool, eshaping a is ic
exp ession and inno a ion. Ins ead o dehumanizing a , his in e ac ion may enhance i , ein o cing
ha c ea i i y has always been shaped by ex e nal in luences—including echnology.
On he o he hand, he ecen "Ghibli e ec " phenomenon, d i en by AI, demons a es how digi al
i ali y can globalize a is ic exp ession, bu i also o ces us o e lec on he e hics and legali y (How
a e ained his AI models? Wi hou pe mission?), and also en i onmen al impac o hese p ocesses,
which equi e as amoun s o da a and ene gy. While we celeb a e AI's abili y o gene a e cap i a ing
Ghibli-s yle images, we mus ask ou sel es he eal cos o his accele a ed c ea i i y: a e we eplacing
human e o and a is ic soul wi h unsus ainable ene gy consump ion o he global ci cus?
Jus as a is s ha e long d awn inspi a ion om o he humans, machines may now play a simila ole. AI
may no be c ea i e in he human sense, bu i is poised o become an in aluable ool in he c ea i e
p ocess, expanding and ede ining ou unde s anding o c ea i i y (Wings öm e al., 2024) and
eshaping bo h a is ic p oduc ion and educa ion.
An oni He nández-Fe nández
Ins i u de Ciències de l’Educació
Uni e si a Poli ècnica de Ca alunya
Re e ences
Chiang, T. (2024). Why A.I. Isn’ Going o Make A . New Yo ke , 31 h Augus , 2024.
h ps://www.newyo ke .com/cul u e/ he-weekend-essay/why-ai-isn -going- o-make-a
C aw o d, K. (2021). A las o AI: powe , poli ics, and he plane a y cos s o a i icial in elligence.
New Ha en: Yale Uni e si y P ess.
Diéguez, A. (2024). Pensa la ecnología. Ba celona: Shackle on books.
He nández-Fe nández, A. (ed.) (2023). C ea i i a digi al. Ba celona: Inicia i a Digi al Poli ècnica,
colecció Diàlegs UPCA s. h ps://upcommons.upc.edu/handle/2117/395833
He nández Fe nández, A. (2025). In eligencia A i icial y diseño. Re is a G a ica, en p emsa.
He nández Fe nández, A. i Fe e Cancho, R. (2023). Lingüís ica quan i a i a i lleis lingüís iques:
de la lingüís ica a la in el·ligència a i icial i la ecnoè ica. Te minàlia, 27.
h ps://doi.o g/10.2436/20.2503.01.190
Holmes, W., Bialik, M., & Fadel, C. (2019). A i icial in elligence in educa ion: p omises and
implica ions o eaching and lea ning. Cen e o Cu iculum Redesign.
Wagensbe g, J. (2017). Teo ía de la c ea i idad. Ba celona: Tusque s.
Wings öm, R., Hau ala, J. y Lundman, R. (2024). Rede ining c ea i i y in he e a o AI?
Pe spec i es o compu e scien is s and new media a is s. C ea i i y Resea ch Jou nal, 36(2), pp.
177-193.
