Noam Chomsky on ChatGPT: It’s “Basically High-Tech Plagiarism” and “a Way of Avoiding Learning”

Chat­G­PT, the sys­tem that under­stands nat­ur­al lan­guage and responds in kind, has caused a sen­sa­tion since its launch less than three months ago. If you’ve tried it out, you’ll sure­ly have won­dered what it will soon rev­o­lu­tion­ize — or, as the case may be, what it will destroy. Among Chat­G­P­T’s first vic­tims, holds one now-com­mon view, will be a form of writ­ing that gen­er­a­tions have grown up prac­tic­ing through­out their edu­ca­tion. “The essay, in par­tic­u­lar the under­grad­u­ate essay, has been the cen­ter of human­is­tic ped­a­gogy for gen­er­a­tions,” writes Stephen Marche in The Atlantic. “It is the way we teach chil­dren how to research, think, and write. That entire tra­di­tion is about to be dis­rupt­ed from the ground up.”

If Chat­G­PT becomes able instan­ta­neous­ly to whip up a plau­si­ble-sound­ing aca­d­e­m­ic essay on any giv­en top­ic, what future could there be for the aca­d­e­m­ic essay itself? The host of YouTube chan­nel EduK­itchen puts more or less that very ques­tion to Noam Chom­sky — a thinker who can be relied upon for views on edu­ca­tion — in the new inter­view above. “For years there have been pro­grams that have helped pro­fes­sors detect pla­gia­rized essays,” Chom­sky says. “Now it’s going to be more dif­fi­cult, because it’s eas­i­er to pla­gia­rize. But that’s about the only con­tri­bu­tion to edu­ca­tion that I can think of.” He does admit that Chat­G­PT-style sys­tems “may have some val­ue for some­thing,” but “it’s not obvi­ous what.”

As the rel­e­vant tech­nol­o­gy now stands, Chom­sky sees the use of Chat­G­PT as “basi­cal­ly high-tech pla­gia­rism” and “a way of avoid­ing learn­ing.” He likens its rise to that of the smart­phone: many stu­dents “sit there hav­ing a chat with some­body on their iPhone. One way to deal with that is to ban iPhones; anoth­er way to do it is to make the class inter­est­ing.” That stu­dents instinc­tive­ly employ high tech­nol­o­gy to avoid learn­ing is “a sign that the edu­ca­tion­al sys­tem is fail­ing.” If it “has no appeal to stu­dents, does­n’t inter­est them, does­n’t chal­lenge them, does­n’t make them want to learn, they’ll find ways out,” just as he him­self did when he bor­rowed a friend’s notes to pass a dull col­lege chem­istry class with­out attend­ing it back in 1945.

After spend­ing most of his career teach­ing at MIT, Chom­sky retired in 2002 to become a full-time pub­lic intel­lec­tu­al. The Uni­ver­si­ty of Hous­ton’s Robert Zaret­sky, who still teach­es, recent­ly offered his own, grim­mer take on Chat­G­PT and edu­ca­tion. “The col­lege essay died years ago,” he argues. “It’s a mug’s game in which a stu­dent sends me an elec­tron­ic file that, when open, spills out a jum­ble of words that the sender pro­pounds to be a fin­ished paper” — to which, pre­sum­ably, the out­put of a machine-learn­ing sys­tem would actu­al­ly be far prefer­able. Most tech­no­log­i­cal “dis­rup­tions” leave both pos­i­tive and neg­a­tive effects in their wake. If the col­lege essay is indeed unsal­vage­able, per­haps Chat­G­PT will final­ly bring about its replace­ment with some­thing more inter­est­ing.

Update: Chom­sky has co-authored an op-ed in The New York Times called “The False Promise of Chat­G­PT”. Find it here.

Relat­ed con­tent:

Noam Chom­sky Explains Where Arti­fi­cial Intel­li­gence Went Wrong

Chat­G­PT Writes a Song in the Style of Nick Cave–and Nick Cave Calls it “a Grotesque Mock­ery of What It Is to Be Human”

Noam Chom­sky Defines What It Means to Be a Tru­ly Edu­cat­ed Per­son

Thanks to Arti­fi­cial Intel­li­gence, You Can Now Chat with His­tor­i­cal Fig­ures: Shake­speare, Ein­stein, Austen, Socrates & More

Noam Chom­sky Spells Out the Pur­pose of Edu­ca­tion

Based in Seoul, Col­in Marshall writes and broad­casts on cities, lan­guage, and cul­ture. His projects include the Sub­stack newslet­ter Books on Cities, the book The State­less City: a Walk through 21st-Cen­tu­ry Los Ange­les and the video series The City in Cin­e­ma. Fol­low him on Twit­ter at @colinmarshall or on Face­book.

by | Permalink | Comments (54) |

Sup­port Open Cul­ture

We’re hop­ing to rely on our loy­al read­ers rather than errat­ic ads. To sup­port Open Cul­ture’s edu­ca­tion­al mis­sion, please con­sid­er mak­ing a dona­tion. We accept Pay­Pal, Ven­mo (@openculture), Patre­on and Cryp­to! Please find all options here. We thank you!

Comments (54)
You can skip to the end and leave a response. Pinging is currently not allowed.
  • naren kini says:

    Too often ask­ing octa­ge­nar­i­ans to opine on tech­nol­o­gy leads to a pre­dictable result. They frame it based on their world view which is now obso­lete. That is under­stand­able. What is not is why any­one would pub­lish their response.

  • Dennis says:

    Seems like the rem­e­dy could be replac­ing essays with face to face ques­tion and answer ses­sions. It shouldn’t take very long to deter­mine who has actu­al­ly learned the mate­r­i­al in ques­tion. Per­haps chat­g­pt gets me through col­lege, but I’m not sure how it helps me dur­ing a face to face job inter­view.

  • garland says:

    this just in, cross­ing an arbi­trary age thresh­old instills one with a ‘world view’ that inval­i­dates all opin­ion.

    con­sid­er­ing that this very obvi­ous opin­ion re: this tech is shared by rep­re­sen­ta­tives of any gen­er­a­tion, it would fol­low that all ‘world views” are obso­lete based on your log­ic. your shod­dy attempt at argu­ing a point only sup­ports the named speak­er’s con­tention that learn­ing avoid­ance is a per­sis­tent trend.

  • Joe says:

    It’s frickin’ Chom­sky you’re ask­ing, not a grandad that does­n’t under­stand smart­phones.

  • Jon Snowden says:

    Real­ly hoped good’ol Noam gets the begin­ning of this “new age”. Appar­ent­ly not.

    Does­n’t mat­ter, he’s been right a lot of times in the past, which is enough for me.

    True hero of our time.

  • Michael says:

    I agree with him ful­ly and I’m young and have heard Chom­sky speak on this sub­ject. He’s more aware of the ins and outs of these sys­tems than the tech jour­nal­ists that hype them up as some­thing that they’re not.

  • Ben says:

    How nuanced and pro­fet­ic smh. Over­rat­ed sci­en­tist

  • Adam says:

    This is what he’s cho­sen to say about it as Amer­i­ca’s lead­ing expert on lin­guis­tics and sci­en­tif­ic method­ol­o­gy.

  • Peter says:

    It is great that Chom­sky came out with this argu­ment, and warn­ing. Yeah, there have been a lot of crit­i­cisms of Chat GPT the last 2+ months, but giv­en that we live in an age when “tech­nol­o­gy”, for all the good it does for us, occu­pies a kin­da meta­phys­i­cal role in the world, as if every new tech thing is good sim­ply because it’s new and tech­no­log­i­cal. Nice to know that the man who gave us “gen­er­a­tive gram­mar”, one of the most impor­tant lin­guis­tic insights, weigh in on “gen­er­a­tive AI”!

  • Arek says:

    Chom­sky him­self is a poor learn­er. He has been a con­trar­i­an revi­sion­ist in regards to con­flicts in Cam­bo­dia, Bosnia, Rwan­da, Syr­ia and Ukraine.

  • Gary Lawson says:

    Why did the inter­view­er waste the first five min­utes talk­ing about his hair­cut, inter­view rou­tine etc.? Annoy­ing.

  • Maarten says:

    Non­sense, just anoth­er old fos­sil stuck in the past. Tools like chat­g­pt are a bet­ter teacher than any human teacher I’ve ever had.. It’s free, quick, knows more and has 100% focus on me and my question(s) and explains things very well. I think peo­ple like gnome should real­ly try to use it first, it’ll make the entire edu­ca­tion sys­tem obso­lete, bar­ring a few very spe­cif­ic stud­ies.

  • Yanga says:

    Stoneage did not end because we ran out of stones, humans sim­ple found a new and bet­ter way of doing things. May be chat­G­PT, and AI in gen­er­al is just a new way of doing things(including learn­ing). We just need to adapt to its devel­op­ment, some­thing humans have been doing through­out the ages.

  • BarrioInvisible says:

    One of the most bril­liant minds in his­to­ry isn’t worth lis­ten­ing to? Ok, bro.

  • Davide Bannerman Sinclair says:

    Look for­ward to read­ing all your pub­lished works in the very near future! Go, bro, you got this!

  • Dennis L says:

    Like any tool it’s val­ue is in how it’s applied. And there is no way of stop­ping it. Those who will use it as a crutch will be like those who can’t do math but can plug num­bers into a cal­cu­la­tor. How much impact that lack of under­stand­ing core prin­ci­ples of dis­cern­ing best sources, the prop­er means of mus­ter­ing an argu­ment, rhetor­i­cal strate­gies, the prop­er fash­ion­ing of lan­guage to con­tent will harm them in terms of career or gen­er­al social inter­ac­tion may be min­i­mal. In fact, a facil­i­ty at lever­ag­ing these AI based tools may help them excel in a sim­pli­fied and sim­ple-mind­ed trans­ac­tion­al cul­ture. And there are good uses for Chat­G­PT, espe­cial­ly for those with the dis­cern­ment to point it at the right prob­lems and to assess the results. But the big­ger impact will be to the sim­ple nature of being human. With­out strength in the skills old-fash­ioned essay writ­ing cre­at­ed in stu­dents (who took on the actu­al task and did­n’t use low-tech tools to bypass all of the above), those stu­dents are reduc­ing their human capac­i­ty for rea­son, inves­ti­ga­tion, under­stand­ing, and the com­mu­ni­ca­tion of their thoughts and feel­ings. That may be a shame, it may even be a death knell for our cul­ture, but it will be part of the future if these tools are not prop­er­ly con­trolled and uti­lized.

  • Edwin Chiumya says:

    Well, when they were first intro­duced, same thing was said about cal­cu­la­tors, google and schol­ar search engines and com­put­ers gen­er­al­ly, that they would make brick and mor­tar libraries extinct. New­er more effi­cient ways of get­ting infor­ma­tion have nev­er replaced a col­lege edu­ca­tion if only because essays only con­sti­tute a small part of the over­all grade for most class­es. Besides, stu­dents still have to con­tend with attend­ing class for at least four years before grad­u­at­ing and will come away with some much need­ed crit­i­cal think­ing skills imper­a­tive to suc­ceed­ing in dif­fer­ent careers. I think we will be fine and the sky is not falling just yet.

  • Dilawar Habibi says:

    Is it free? It wouldn’t let me do things I want­ed, unless I pay $19.99 per month.
    Can you tell us how do get it for free?

  • Efrosini Drimoussis says:

    I have heard that ope­nAI will retain all mate­r­i­al pro­duced and some­how make it avail­able to schools/institutions to scan for pla­gia­rism. Not sure if accu­rate and/or pos­si­ble, but I am opti­mistic edu­ca­tors will find a way to deal with this prob­lem.
    As to lazy learn­ers, we are not all as clever as Chom­sky and tools like AI may help us accom­plish more. I have two uni­ver­si­ty degrees (includ­ing engi­neer­ing MSc), so while not lazy, I’m also not bril­liant, and I feel this may help me learn/do more.

  • peter stanley says:

    chat­G­PT and the hype is not wor­thy. It does­nt inter­face with the inter­net, just some ency­clo­pe­dia. It does­nt know how to react with abuse or sex­u­al atten­tion. Even Dr Sbait­so was more fun. At least it kbows the dif­fer­ence between g code and g spot. If i were an instruc­tor prompt­ing essays, i would see right thru kids using chat­g­pt. Just make them write the essay on paper. Let the best cal­lig­ra­phy win lol

  • Neil Krane. says:

    Ageist twerp!

    Imbe­cile, Chom­sky is a nona­ge­nar­i­an.

  • Justin says:

    Agree on all points.

  • Glenn Montague says:

    This is 100% cor­rect. And the rea­son it is done is because the answer the inter­view­er knows they will get sup­ports the inter­view­ers agen­da.

  • Sowmya Jess says:

    The man puts it bril­liant­ly! His age defy­ing brain is capa­ble of an aston­ish­ing astute­ness that neu­ro­sci­en­tists should try to unrav­el. I have tried Chat GPT. It can pro­duce clev­er­ly craft­ed pla­gia­rism at the speed of light and it will take an intel­li­gent mind to detect the lack of orig­i­nal­i­ty. It can nev­er come up with a “ Crime snd Pun­ish­ment “ or a “ Mill on the Floss” or an “ Inno­cents abroad”. It can write a cheap thriller but it will nev­er pro­duce a great “ orig­i­nal “ lit­er­ary work. With stu­dents resort­ing to this to write essays, human cre­ativ­i­ty will ebb and die. What it has done is per­fect the art of human pla­gia­rism.

  • Vijay says:

    Old gen search engine pro­grams are infor­ma­tion pro­cess­ing sys­tem.

    Chat­g­pt is car­ry­ing huge bur­den as it is from intel­li­gence fam­i­ly.

    Each step of pager­ank or oth­er sim­i­lar tech­nol­o­gy was con­trolled and guid­ed by pro­gram­mer, sys­tem design­er. That’s domain spe­cif­ic intel­li­gence poured in by human. Chat­g­pt /current gen intel­li­gence mod­els are based on gen­er­al pur­pose prob­lem solver evolved in last 80 yrs, but again that intel­li­gence is poured in by human.

    Pager­ank can not be called pla­gia­rism, but if some­one claims self intel­li­gence, it would face ques­tion of pla­gia­rism.

    Human pre­ferred to use machine from cen­turies. And defi­nate­ly it lim­its our learn­ing in some area. Some­how this crazy crea­ture bounce back with new way of learn­ing. It’s same as I lost ener­gy to walk miles because of horse and then aero­plane.

  • MrPrentiss says:

    I think it’s pla­gia­rism in much the same way as using a cal­cu­la­tor is. No one gets upset any­more if you use a cal­cu­la­tor. It did­n’t cause the end of math. Edu­ca­tors that get out in front of this tech­nol­o­gy and teach it in the class­room are the only right thinkers. Peo­ple who want to fig­ure out ways to try to ban it are myopic and reac­tionary. Embrace it, teach with it. It’s not going any­where, nei­ther is the cal­cu­la­tor.

  • Michael says:

    A few thoughts:

    It’s clear Chom­sky has not attend­ed a col­lege course as a stu­dent in a long time. He seems unaware at how infre­quent an essay is request­ed in many fields. I can count on one hand the num­ber of essays I wrote in col­lege, and that was in the 90s. Essays are only applic­a­ble in a few social stud­ies that don’t gen­er­al­ly have action­able tasks (e.g. regur­gi­tate knowl­edge rather than solve this prob­lem).

    There’s no sense in per­fect­ing skills which will be obso­lete. Just like folks are say­ing with cal­cu­la­tors. In the future, no one is going to pro­fes­sion­al­ly write an essay: it takes way too long and like­ly isn’t as well under­stood by oth­ers as you think it is. Your first draft will always be a chat AI.

    Peo­ple who thrive, includ­ing teach­ers, will be the ones who embrace change. This is true with cal­cu­la­tors, then computers/laptops; typ­ing over hand­writ­ing; open notes over closed book; and soon to be, chat AI over essays. It’s not good or bad: it just is.

    It’s true that it’s hard to teach an old dog new tricks. I’m already feel­ing that myself and I’m only half his age.

  • mschribr says:

    Chom­sky is not think­ing clear­ly. It’s only pla­gia­rism or cheat­ing if I pass it off as my own. He says it’s not obvi­ous what val­ue chat gpt has. How about being the ulti­mate teacher? Chat gpt teach­es at your lev­el. It will teach in any for­mat you ask, a poem, pirate, gang­ster or Shake­speare. It knows every­thing. Is always avail­able and patient. Pla­gia­rism has always been a prob­lem. A stu­dent can get his friend to write the essay. If you want stu­dents to write essays, then write it on a test, not a home­work.

  • RasmusR says:

    Chat­G­PT does­n’t know what’s true. It goes off what it thinks sounds like lan­guage in the request­ed for­mat. It fre­quent­ly claimed to me that 3+4=6.

    It has to be restrained due to the amount of mis­in­for­ma­tion avail­able on the inter­net. It would spout anti­se­mit­ic non­sense when asked about the jews, because nobody writes more about the jews than the anti­semites.

  • Tom McLernon says:

    I agree with the assess­ment. It is very good at regur­gat­ing mate­r­i­al, but using that mate­r­i­al to log­i­cal­ly deduct some new dis­cov­ery or pos­tu­la­tion not so good.

  • David says:

    A cal­cu­la­tor does­n’t pre­vent you from learn­ing cal­cu­lus

  • Chris Joseph says:

    NoamG­PT is a trust­ed expert sys­tem based on a bio­log­i­cal plat­form, and near­ly a cen­tu­ry of train­ing data

  • Daniel says:

    Hon­est­ly, any aca­d­e­m­ic who’s fooled by a Chat­G­PT essay isn’t qual­i­fied to teach. Like that song in the “style” of Nick Cave, what pass­es for a Chat­G­PT essay reads stilt­ed and half-baked, is full of clichés and bereft of actu­al devel­oped orig­i­nal ideas. And it usu­al­ly just gets it wrong. If a stu­dent pass­es in an essay sus­pect­ed to be writ­ten by AI can’t you just test its legit­i­ma­cy by ask­ing them to explain their argu­ment? I would­n’t wor­ry that the sky is falling just yet.

  • Tim says:

    It’s a rel­a­tive ide­ol­o­gy to label Chat­G­PT in this man­ner, as for every­one it’s dif­fer­ent with their unique back­grounds.

    For me Chat­G­PT is…

    1.A teacher that will NEVER get mad at you no mat­ter how many ques­tions you ask.

    2.A teacher I CAN ask ANYTHING at ANYTIME and get an in-depth response because they are NOT going through a divorce, an alco­holic, caught up in scan­dals, doing drugs, hun­gover, have events to attend and a sub­sti­tute teacher NOT qual­i­fied in the field and try­ing to teach you when all they are doing is basi­cal­ly babysit­ting you until your teacher returns. ALSO a teacher who is’t out­dat­ed in their field as with exten­sions Chat­G­PT can look at updat­ed inter­net data in sec­onds rather than hav­ing to take recer­ti­fi­ca­tion class­es to try and catch up.

    3.A teacher that nour­ish­es ones learn­ing style and does NOT crit­i­cize it.

    4.A teacher who helps is lit­er­al­ly an intro­verts best friend.

    5.A teacher who can one on one brain­storm with you on ANY sub­ject and won’t say “That’s out of my field, as it can access 1,000s of dif­fer­ent fields.

    6.A teacher that does NOT treat stu­dents like #$%&, unlike real teach­ers that have tenure. So Chat­G­PT is always kind a cour­te­ous. (Basi­cal­ly for any­one that goes to crit­i­cize my mes­sage here or any typos I do. And no Chat­G­PT isn’t writ­ing my mes­sage to all the haters.)

    7.A teacher who has taught my child more than I ever could in help­ing him progress even fur­ther because we home school and he’s learn­ing already at a high school lev­el at age 8, but Chat­G­PT is help­ing us push his aca­d­e­mics in an even bet­ter way.

    8.The list is end­less…

    And I’m not hat­ing, but it seems like it’s a sit­u­a­tion where teach­ers are scared of being replaced. Well, they will and should be. I still remem­ber tot his day of my 3rd grade teacher com­ing in, from what I know now was a hang­over, and telling all the chil­dren, again chil­dren to “Shut the #$%& up.”. I almost failed her class, actu­al­ly all x30+ stu­dents were all almost held back that year, Well if AI takes over, #$%& teach­ers won’t exist any­more. Every­one will have an equal oppor­tu­ni­ty to learn one on one. I’m for that.

  • James says:

    I don’t agree, I think it’s a great resource FOR learn­ing. I’ve learned a lot since I’ve start­ed using it so to say it’s some­thing that avoids learn­ing is kin­da ridicu­lous to me

  • Robert McGee says:

    I’m old enough to remem­ber when TV was new and the same dis­cus­sions were tak­ing place. Argu­ments ranged from TV ruin­ing soci­ety to being a high pow­ered teach­ing tool that would pro­pel us into the future. This hap­pened again with the inter­net. I’m ful­ly aware of the evils of today’s tech­nol­o­gy, but I wouldn’t want to go back to a world like it was before TV. Thus, I’m fas­ci­nat­ed and excit­ed about AI, but ful­ly expect it to unleash soci­etal hor­rors we can’t even imag­ine yet.

  • Sirpnie says:

    Of course AI isn’t going go to be real­ly pos­si­ble to reg­u­late. But it can help with learn­ing. Just like how the print­ing press made it so books were viable. To libraries. To TV. To many oth­er things. Essays? Who care? Well not gov­ern­ment employ­ees hav­ing to make hun­dreds on rel­a­tive­ly cus­tom emails that they can have AI write then they proof read it. Take a hour plus thing to a few min­utes. If it comes to sci­ence, peer review. Prob­lem solved. If it’s pure­ly for teach­ing crit­i­cal think­ing, well just find new­er and bet­ter ways that can use Ai togeth­er. There’s more to thought than writ­ing an essay.

    I can see the rea­sons for wor­ry though. Like what do you do when poems are best writ­ten by AI? We aren’t there yet sure. what about when AI is all you need to make games? It’s an under­stand­able fear to see your pas­sion be at risk. It will be eas­i­er to do all sorts of immoral things. But just like with most tech, the use deter­mines the moral­i­ty. Hope­ful­ly it remains most­ly a tool. It’s still at the tool stage.

    We hope­ful­ly aren’t close to sen­tient AI because then that’s a whole new ball­game where you got­ta fig­ure out rights for AI because, well, they would be able to feel and expe­ri­ence the cru­el things done to them in the way we do. Humans already strug­gle with shut­ting down things that resem­bles a sen­tient AI let alone a real one that you can actu­al­ly com­mu­ni­cate with.

  • GLR says:

    Chat­G­PT makes life easy and also leads to ear­ly demen­tia and ear­ly death Your choice bros who go with chat­G­PT

  • George says:

    Oh, look, every­body.

    It’s the abject, dis­gust­ing sovi­et immi­grant who can’t stop crit­i­cis­ing the same regime his fam­i­ly walked on corpses and sold their souls to be a part of and live under!
    Cham­pagne social­ism tru­ly bev­er gets old.

    But wait, it gets bet­ter! the hard­core social­ist is now telling us about his con­ser­v­a­tive opin­ions regard­ing study­ing.

    Tru­ly, many peo­ple are good liars and abject oppor­tunists, but no one holds a can­dle to the Rus­sians.

  • Clive Appleby says:

    Edu­ca­tion is a chal­lenge to many. This is why ai ethics are a con­cern. Thank­ful­ly the busi­ness ethics white paper is near­ing its com­ple­tion and some inter­est­ed indus­tries have begun to address this con­cern from the soft­ware e and hard­ware indus­tries.

    Most peo­ple won’t notice the dif­fer­ence but Gram­mer­ly alone already has some ethics in its gram­mar tools Next is MS Office, and the nice folks at Google and essen­tial­ly every­one in tech will get the option to add an indus­try stan­dard eth­nics mod­ule to their soft­ware and hard­ware.

    It’s about find­ing a mid­dle ground with anti-pla­gia­rism at the core and a whole host of oth­er sec­tors like edu­ca­tion sys­tems.

  • Jannie Pretorius says:

    If you steal from one per­son, it’s called theft. If you steal from many peo­ple, and men­tion their names, it’s called research. Ha!

  • mschribr says:

    Ras­musR, I have been using chat­g­pt. I have not seen those prob­lems. It is still in the test­ing phase. If it does have these prob­lems when the test­ing phase is over than peo­ple will use some­thing else.

  • Wokkel says:

    Ask that to one of my kids…

  • Themerc says:

    If Chat­G­pt is a way to stop learn­ing, that would mean writ­ing essays and mem­o­riza­tion is learn­ing. Which it isn’t.

  • Youssef TIRIZITE says:

    I think we can also use it to aug­ment learn­ing. The prob­lem lies in our eval­u­a­tion meth­ods, not tech­nol­o­gy.

  • Euyre says:

    It’s obvi­ous that Chat­G­PT does­n’t under­stand any­thing it’s say­ing. Lan­guage that is devoid of expe­ri­en­tial ref­er­ence has no mean­ing — it’s just cir­cu­lar math.

    If a dark “mind” only knows that bleeph leads to bramph and dorhskj, it knows noth­ing. Lan­guage with mean­ing to its hear­ers and speak­ers REQUIRES a world exter­nal to it, a world of objects and dynam­ics that can be rep­re­sent­ed by objects and dynam­ics of lan­guage. A sym­bol needs a ref­er­ent, or it is not a sym­bol.

    Tru­ly intel­li­gent AI will have to be able in some way to have an expe­ri­ence of the world we share. It must per­ceive sim­i­lar inputs that humans per­ceive in order to speak cre­ative­ly (sym­bol­i­cal­ly) with humans, rather than mere­ly “speak” “pre­dic­tive­ly” (as in blind­ly math­e­mat­i­cal­ly).

  • Richard Kelly says:

    Too often igno­rant young peo­ple offer opinions/responses to opin­ions with data/understanding insuf­fi­cient to mer­it valu­able opinion/thought — this is made worse when non-stan­dard­/un­com­mon words are used to reinforce…and they’re the wrong words.

    Case in point? Noam Chom­sky is 94, which makes him a nona­ge­nar­i­an.

    To judge his opin­ion to be with­out val­ue, you should/must under­stand it and eval­u­ate it on it’s mer­its — or you could have a ham-hand­ed attempt at an ad hominem attack, which dimin­ish­es any mer­it your com­ments or case may have had even fur­ther.

    Try read­ing or lis­ten­ing to some of his com­ments with an open mind — he’s stud­ied the sub­jects of con­scious­ness and human intel­li­gence, among oth­ers for over 70 years — it’s very fool­ish to throw his opin­ion out undi­gest­ed.

  • skyman mentalist says:

    Hope­ful­ly you won’t stop the flow of such mag­i­cal mate­r­i­al!

  • Sve7en says:

    He’s 94, not 84.
    We’ve got stop incor­rect­ly label­ing peo­ple as ‘old’ then dis­miss­ing their thoughts. When some­one shares their thoughts, they aren’t edicts one must fol­low — they are con­tri­bu­tions to a deep­er con­ver­sa­tion.

  • Guy Lewis says:


  • ahmad ali says:


    Good day!

    I’m reach­ing out to check if you are accept­ing con­tent from guest con­trib­u­tors on your blog ( ) . We are try­ing to estab­lish our­selves as a trust­wor­thy brand and we think the best way to do that is through qual­i­ty con­tent.

    So, if you accept blog posts on your web­sites, share the details, please.


  • smart@ss says:

    I dis­agree. I’ve been famil­iar with his ideas for for decades and even back in the 1990’s his the­o­ries were tired and going nowhere. I would­n’t have advised any­body to go into the “field” of Lin­guis­tics then and it’s obvi­ous now. Chom­sky’s ideas will nev­er pro­duce any­thing like Google Trans­late. Only peo­ple inter­est­ed in the his­to­ry of AI should look at his stuff.

  • smart@ss says:

    Says Chom­sky,
    “may have some val­ue for some­thing,” but “it’s not obvi­ous what.”

    How about the val­ue of lay­ing ruin to your idea that the human mind cre­ates gram­mat­ic sen­tences because there is a ded­i­cat­ed and inborn struc­ture in the brain? That idea is dis­re­spect­ful to the human mind, as it pre­cludes the abil­i­ty of humans to learn how to cre­ate gram­mat­ic sen­tences. And fur­ther, assert­ing that your hypoth­e­sized inborn struc­ture is the only unique qual­i­ty to humans among ani­mals, low­ers human dig­ni­ty and is less than what I would expect from a “pub­lic intel­lec­tu­al”.

  • Adam Davidson-Harden says:

    Chom­sky has a good point, as usu­al — as an edu­ca­tor myself I can attest to it. Most of the aca­d­e­m­ic integri­ty issues I face are now AI-based, and I’m def­i­nite­ly con­cerned about stu­dents side-step­ping lan­guage and writ­ing skills due to the fact that AI-pow­ered skills can essen­tial­ly write for them, or trans­form poor writ­ing into tech­ni­cal­ly good writ­ing. He is now 90, and still very much lucid, it’s spelled ‘octo’, not ‘octa’.

Leave a Reply

Open Culture was founded by Dan Colman.