What do sandwiches have to do with AI? Data reformist Kasia Chmielinski helps us think about artificial intelligence with a useful food metaphor — and breaks down why AI systems should have “nutrition labels” to ensure the development of fairer, more transparent algorithms.
If you love watching TED Talks like this one, become a TED Member to support our mission of spreading ideas:
Follow TED!
X:
Instagram:
Facebook:
LinkedIn:
TikTok:
The TED Talks channel features talks, performances and original series from the world’s leading thinkers and doers. Subscribe to our channel for videos on Technology, Entertainment and Design — plus science, business, global issues, the arts and more. Visit to get our entire library of TED Talks, transcripts, translations, personalized talk recommendations and more.
Watch more:
TED’s videos may be used for non-commercial purposes under a Creative Commons License, Attribution–Non Commercial–No Derivatives (or the CC BY – NC – ND 4.0 International) and in accordance with our TED Talks Usage Policy: . For more information on using TED for commercial purposes (e.g. employee learning, in a film or online course), please submit a Media Request at
#TED #TEDTalks
@TimCCambridge
June 14, 2024 at 7:07 am
~Hi, AI.
Now the ” virus ” is loose, tell everyone how positive you are with your gain-of-function mindset.
@educationiskey3331
June 14, 2024 at 7:22 am
Recipe for AI?
Racism. Because alot of companies have Racism belt into it.
@TesserId
June 14, 2024 at 8:03 am
Funny typo. I think your spell checker let you down.
@toni4729
June 14, 2024 at 7:23 am
The obvious think regarding food is to only buy things that don’t need labels. I do. It doesn’t come in a box, only one item at a time. A vegetable, a fish, a meat. Eggs. Simple isn’t it?
@toni4729
June 14, 2024 at 7:27 am
Unfortunately the government make too much money out of big business to stop all the rip-offs.
@toni4729
June 14, 2024 at 7:30 am
We didn’t have this problem when food was food and supermarkets didn’t exist. You’re not old enough to remember those days. When there was a butcher and a greengrocer on every corner and we had to go to the dairy to get the rest. That’s all we needed. No chemicals.
@handlewastaken
June 14, 2024 at 8:37 am
Yes. We didn’t have this problem either back when computer programs were simple calculations that you could look at instead of an uncountable amount of matrix multiplications being done per second.
Also, there’s a reason supermarket food took over – it’s cheaper, safer in the short term (as in: you don’t have to worry about getting poisoned that often), and easier for consumers.
@rikachiu
June 15, 2024 at 12:09 pm
Contrary to this belief, it’s WAY better now than it ever was back then by all metrics.
@toni4729
June 15, 2024 at 12:24 pm
@@rikachiu People are being poisoned now.
@toni4729
June 15, 2024 at 12:26 pm
@@handlewastaken You have a short memory. We were not poisoned then, we shopped for fresh food daily.
@JoshPitts530
June 14, 2024 at 7:31 am
Too bad the AI safety board is comprised of tech conglomerates
@ShahrezadNorMohammadiy9116
June 14, 2024 at 7:32 am
Success usually comes to those who are too busy to be looking for it.
@mehmet.albyrk
June 14, 2024 at 7:36 am
Talking too fast 😮
@michellenm92
June 14, 2024 at 7:50 am
If you click settings, you can lower the playback speed ❤
@mehmet.albyrk
June 14, 2024 at 7:55 am
@@michellenm92 really thanx 👍
@coolParadigms
June 14, 2024 at 7:56 am
A new trend in AI vocabulary : the substitution of neuronal network with algorithms! In reality they are almost as far from algorithms as we are, they are artificial brains.
@TesserId
June 14, 2024 at 8:00 am
Can’t help but think back to my basic statistics class. We learned that there weren’t just measures of central tendency (averages), but that there were measures of variation. I would hope that by now there would be better ways to deal with this.
@TesserId
June 14, 2024 at 8:01 am
BTW, can’t help but think of the old joke about the statistician who would stand in a tub of ice water with their head in a hot oven and say: “on average, I feel fine.”
@TesserId
June 14, 2024 at 8:00 am
Can’t help but think back to my basic statistics class. We learned that there weren’t just measures of central tendency (averages), but that there were measures of variance. I would hope that by now there would be better ways to deal with this.
@andycordy5190
June 14, 2024 at 8:19 am
Brilliantly conceived and presented. Thank you.
I loved the analogues of the food labelling and in particular the gluten free sandwich.
Increasingly, we are seeing we recognise that AI, having been developed for users needs to appeal to users. Safety is a crucial aspect of that reassurance and if providers expect users to return again and again to their platform in a competitive market, they will offer that reassurance freely.
Making it verifiable is a big task but recognising it is essential is a big step.
@TesserId
June 14, 2024 at 8:33 am
Food labeling has had it’s history of being molested by lobbyists. I remember details on certain products (some famous ones) that at some point got rolled up into a catch-all term like “seasonings” and “spices”. Presumably, this was to protect proprietary recipes. AI corporations are going to (or already do) have their lobbyists. I just don’t feel optimistic about this.
@jmillart
June 14, 2024 at 9:08 am
AI needs to be open and honest about WHERE it gets all of its training data. Most art AI has a ton of theft and pays no royalties it should to the original artists.
@vesawuoristo4162
June 14, 2024 at 9:24 am
Great talk
@sabong8903
June 14, 2024 at 9:33 am
Ted, show us how smart and forward thinking you really are and refer to AI as ai.
@mfbe73
June 14, 2024 at 9:42 am
Has anyone begun to grasp that perhaps they are way underqualified to make prescriptions about AI? At least until there is a theory of what causes phase changes in neural nets!
@BrainiousPodcast
June 14, 2024 at 9:49 am
So curious about the future of Ai! We also just launched our channel and plan to make videos of it, althought it advances too fast to feel like knowing enough.😅😅
@BreezyVIL
June 14, 2024 at 10:19 am
Blue eyed Samurai lookin ah
@BUY_YOUTUB_VIEWS_195
June 14, 2024 at 10:46 am
I’m hooked on your channel
@lonewolf36s
June 14, 2024 at 11:21 am
Since we cannot remove sociopaths from this planet, we therefore cannot remove sociopathic trained AI.
This will never be fixed. The interests of evil people will always eclipse the interests of the well intentioned.
And even further, we are training all the models on corrupted (biased) data. Because there’s no such thing as unbiased content.
“AI” will always have a dark side. Just like us. For we created it.
@gendispragitaputri8541
June 15, 2024 at 1:34 am
I really support the idea of data nutrition ingredients, It’s very insightful thoughts 👏❣
@RezaDadashi-rl7ut
June 15, 2024 at 6:04 am
Fri 25june2024 12:20
برای زنده ماندن،بایدهایی تعریف شده وجود دارد.این بایدها،بارها در تاریخ مکتوب و قبل از آن تکرار شده است.بایدهایی بسیار زشت و زیبا و حتی بدتر و بهتر از زشتی و زیبایی.یکی از این زیباییها نه،بلکه زیباترین باید،برای زنده ماندن را تعریف میکنم.نام آن جنگ است،بین حیوانات مشهور به نزاع،و نزد گیاهان،نامش رقابت است.منظورم از اصطلاح – بایدها برای زنده ماندن – در معنی یعنی اعمالی که نوع جاندار،برای اینکه بتواند زنده بماند،باید آنها را رعایت کند.جنگ،برترین شرط مطلق برای بقای انسان است و تا بوده همین بوده.از طرف دیگر یک مداد شمعی بنام عشق وجود دارد که آنهم بعنوان یکی از بایدها برای بقای انسان،تعریف شده است.اما عشق بین دو شخص غیر همخون،هیچ معنایی ندارد.و آن یک دم هوس و مابقی عمری زندگی در جهت خواست امیال حیوانی میباشد.در جنگ و عشق همه نوع رفتار قابل توجیه است.در جنگ اگر قافیه را شل بگیری باخته ای.در نوع لوث و شبیه مداد شمعی شرط بقا یعنی عشق.اگر کوتاهی کنی،هویت خودت را برای همیشه باخته ای.در نوع کامل شرط بقا یعنی جنگ،اگر کوچکترین ترحم و کوتاهی مقابل دشمنت انجام دهی.این یعنی،فی الفرض – صیغه شده است دیگر – جاری شده،پس: در ابتدای محور تلاش برای بقا و در جنگ مقابل دشمنت – مغلوب – میشوی.و در انتهای این بردار که محل تجمع بهترین احتمالات میباشد،دچار زخم دشنه،گلوله و یا کینه میشوی.و در نهایت مجبور به بقا،با تعفن این زخم خواهی بود.در همیشه تاریخ،برندگان مطلق جنگها،آنهایی بوده اند که چشم و گوش و دهان خود را دوخته و فقط کشته اند.چرا جنگ؟!؟ مصدر کوچه بازاری فعل گرفتم یا – گیرم – را بر این بنا میکنم که؛ شخصی با توانایی کامل و نیروی مالی،فکری،علمی،نظامی،نوآوری و اجرایی؛پدید آمده و با تمام ارکان شهود ادراکی و عقلانی،اثبات میکند که میتواند.آنگاه اعلام میکند: ای مردم،ما میخواهیم شما را به امنیت مالی و جانی،ثروت،رفاه،تأمین،خجستگی،گردش و سفرهای شادمان،شغل کامل و محترم،خانه لوکس و عالی،شهر رؤیایی،فرهنگی پاک ورای تصورها وتجسمهای خودتان تا به امروز،و الی آخر تا شکوه و سعادت در زندگی و بعد از مرگ – همان سکوت و بهت،مابین دو زندگی،یا مات – برسانیم.به نور و تاریکی سوگند که در اکناف هر جا از همه جای زمین،عده ای جمع میشوند و از رجز گرفته تا شعار و ترانه و کارنیوال راه می اندازند که: مرگ بر اینها که میخواهند الف تا یائ الی آخر انجام بدهند تا ما به سعادت برسیم.یعنی بزبان دمروها یا دمر خفتگان ،همیشه و با اینکه تمام انسانها میدانند جمع رقم یک با یک میشود دو.اما باز در صورت محتمل شدن اجرای این معادله ساده و امثال آن،مطلقا مردمانی جمع شده و مخالفت خواهند کرد.حتی مخالفان هم مخالفان تصاعدی خواهند داشت.این بزبان ساده یعنی شرارت.حال،چه شود؟!؟ – جنگ – و بزرگترین چاره برای بشر این است که: حواست باشد،جنگ کن و ادامه بده.و هرگز بر دشمنان افکار صحیح و آزمون شده خودت،ترحم روا نکن.درمان گروه هایی که در بنیه ایشان،مخالفت با حقیقت وجود دارد،فقط مرگ است.زن و مرد و کودک و پیر آنها باید کشته شوند تا ژنتیک و نطفه آنان از زمین پاک شود.دیدن اینکه،در ناحیه وقوع یک درگیری نظامی فوق استراتژیک،فرماندهان و متعاقب آن سربازان،صرفا بدلیل حضور زن و کودک و پیرمرد و پیرزن؛چند دقیقه فقط چند دقیقه،تعلل کرده اند و بعد خودشان که غالب میدان بوده اند،مثل سگ کشته شده اند،بسیار با معنا است.یله چریدن پیرمردان،پیرزنان و کودکان در نواحی درگیری نظامی و ممانعت از اجرای فعل – ترک محل تخاصم و درگیری نظامی – بسیار با معناتر است.یقین دارم،تلاشم برای فقدان تشابه متن با موعظه؛نتیجه داده است.پس ادامه میدهم.شرارت،قبل از آفرینش انسان،حیوان وگیاه؛و زمانی که مخلوقات عبارت از ملائک،جنیان و مخلوقات ناشناس دیگر بودن؛وجود داشت و آفریدگار از آن آگاه بود.او خود میداند چگونه اراده نمود،اما من یقین دارم هدف از آفرینش کالبدهای مختلف،و سرشتن و شکل دادن آن با قدرت اجرای خودش یعنی دستان ملائک وجنیان – فقط و فقط،این بوده و هست که – محمل و محفلی برای ظهور وحضور جسمانی یا همان تجلی شرارت،ایجاد شود.شبطان کرنش نکرد.ملائک کرنش کردند اما عده ای شروع به نجوا بین خود و طرح سؤال دلیل برتری بشر بر خودشان نمودند.آنهم با اعلام تنزیه خودشان از طغیان ! خداوند نیز اراده مکتوم خود را بشکلی که همه ما میدانیم ادامه داد.اما تمام نکات افتراق بین خیر و شر را به تمامی انحائ و در تمامی زمانها،بارها تکرار کرد.سفاهت مخلوق شر و شرارت گستر در اینجاست که او در طول ادوار آفرینش خود به نقطه ای رسیده که هوای فتح و تسخیر ملک پادشاهی مالک هستی خود را دارد. چه سفیهانه 🤬
@badbunny1984
June 15, 2024 at 7:08 am
Terrific talk.
@rikachiu
June 15, 2024 at 12:14 pm
AI doesn’t exist yet. Not until at least AGI (most argue super intelligence until we really get there). And we won’t get to AGI without a crapload of data. So much data that they are creating synthetic data. Even before that, we don’t generate enough actual electrical power to reach AGI currently.
@handlewastaken
June 15, 2024 at 2:45 pm
Yes, and even so it’s difficult — the way the data is processed also has to be improved. The way that, for example, large language models work right now is like asking a student to skim through a bunch of textbooks, many of which are full of junk, and then making that student answer a very hard test. You could get them to do better at the test by giving them even more, higher-quality textbooks, however, the best strategy would be to make the student actually read and reason upon the books. It’s also important to note that the speaker in this video isn’t so much talking about AGI and multimodal conversational models but instead about the specialized systems that are already deployed and making decisions (like who gets access to medical benefits).
@rikachiu
June 15, 2024 at 6:06 pm
@@handlewastaken My argument really was, in order for us to get to ‘true’ AI, it’s going to need as much data as possible and privacy is the antithesis of getting there.
@tiberiousjc3739
June 15, 2024 at 1:41 pm
Garbage in garbage out, AI is doomed to be a buggy pile of junk and will never be able to tell the difference from fact and idealism, or truth from lies. It just runs code, algorithms based on human interaction and input, an over bloated search engine without a thought of its own.
@NandKumar-qq3xk
June 15, 2024 at 10:35 pm
Every Atome and it’s family’s no. Is as equally one to Even Billions and Multyplyings way” charecter of a Atome and Rest’s Atomic structurs and wee having Cofident about one to further rests” and it is material world” and ours Mixing technology, ant Purity and Guarranty pressurs of Addvertisment, Basic is Outdated of today is Baseless ?
@NandKumar-qq3xk
June 15, 2024 at 10:38 pm
Ressistance and matarial science wee facing in Electrical Transmissions way”
@Freewxll_
June 16, 2024 at 6:41 am
Welcome to the age of metadata everyone the age of information is over imo
@reyfox7366
June 16, 2024 at 1:29 pm
Nutrition label? Nope! A. I. needs to not exist, at all! I really don’t like the potential it can achieve in matter of months which takes humans a lifetime. Things will become even more fast-paced for humans in the technology sector. And that is when machines take over. Using a.i. is the biggest mistake humanity will commit!
Giving it our medicals history will only give an a.i. a gun in it’s hand.
We absolutely shud stop anything a.i.
@user-ez2gw4nt3i
June 18, 2024 at 8:02 am
11:00
@ganjei5236
June 19, 2024 at 3:26 pm
Wow, thank you, this makes so much sense. I hope it catches on!