When you need to make a decision, you might turn to AI for advice. But what happens when you let algorithms decide more and more things for you? Computational social scientist Sandra Matz reveals how AI models limit what you see, discover and try, subtly flattening what makes you unique. Hear her tips for how to use AI to fuel your curiosity and expand your world instead. (Recorded at TEDxNewEngland on October 30, 2025)
Join us in person at a TED conference:
Become a TED Member to support our mission:
Subscribe to a TED newsletter:
Follow TED!
Instagram:
LinkedIn:
TikTok:
Facebook:
X:
The TED Talks channel features the best talks and performances from the TED Conference, where the world’s leading thinkers and doers give the talk of their lives in 18 minutes (or less) — plus originals, podcasts and exclusive content. Look for videos on Technology, Entertainment and Design as well as science, business, global issues, the arts and more. Visit for our entire library, transcripts, translations and personalized recommendations.
Watch more:
TED 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 the TED Talks Usage Policy: . For more information on using TED for commercial purposes (e.g. employee learning, in a film or online course), submit a request at
#TED #TEDTalks #AI
@explorerofmind
September 28, 2026 at 11:40 am
We already were basic. Don’t pretend.
@DragolichRL-1
September 28, 2026 at 11:51 am
AI????? Really? How about all the gangs. The people that claim they are helping but are not. This lady is dum. I think people like her are more dangerous. What does she mean AI? Fronteir models? Local? What pipelines? WHat tools? No we talk about ice cream. Please stop letting these chalotons up here for grant money.
@DragolichRL-1
September 28, 2026 at 11:53 am
its the softmax formula built right into the paper “attention is all you need”. This lady is a softmax formula. She only thinks in binary. AI can never be a 0 or 1. Nor is life. She is off in her wavelengths. She is a wall. Not a destination.
@Skyace13
September 28, 2026 at 11:54 am
So this isn’t about AI but engagement algorithms that have been around for years…. Not to say that AI hasn’t accelerated it, just a bit underwhelming.
Edit: glad to hear about the data behind it. Some of these videos present ideas, this was a presentation of research.
@drumbum3.142
September 28, 2026 at 12:03 pm
– Everyone is Different. Some More than Others. Some Less. While there might be occassion(s) to.. say, Believe in the Same Thing or Act in the Same Way; that will Not always happen. That does Not mean though/however that one should neccessarily Abandon certain specific opinions, assertions, and/or beliefs – sometimes just the opposite.
@UseLessFact-ory
September 28, 2026 at 12:14 pm
Either way…
We deserve EVERYTHING that is coming…
for letting people fall to their death..
(We simply allow them to die) – but we can save all of them
The WORLD just decided everyone needs to create their own job…
Enough work for 50% of Country currently.
(The rest slowly die)
MILLIONS already dead… 10s of MILLIONs more to come.
(STILL NO PLAN in 2026)
Yes…We DESERVE EVERYTHING THAT IS COMING for this GENOCIDE.
(Largest in Human HISTORY)
-Homeless US Army Veteran
@homewall724
September 28, 2026 at 12:25 pm
Public schools and central planning rulers have been doing this for a long time, and before that, it was religion. You can’t have real tolerance or diversity without first preferring free people in free markets with free speech and association because a voluntary society promotes variety while central planning and services makes us all slave-cogs.
@dave23024
September 28, 2026 at 12:28 pm
The empowerment con was turning everybody into the same person. That was the goal, that was how it went in practice, and everybody got stuck with halved wages.
@J.Tronix
September 28, 2026 at 12:28 pm
Tell me you live in an echo chamber without telling me you live in an echo chamber
@JamesOfEarth
September 28, 2026 at 12:38 pm
Or you could just not use it…
@AwesomeCrackDealer
September 28, 2026 at 1:29 pm
just turn up the temperature on the token generation and you will introduce more “exploration” into the explore/exploit mix. This talk comes from someone that does not understand what she is criticizing on a first derivative level – and it is kind of weird that TED allowed this on stage
@craig-warman
September 28, 2026 at 1:31 pm
Garbage. Problem : “AI is making us basic”. Solution: AI will now also regurgitate your secondary and tertiary preferences making you appear super unique.
@TomNook.
September 28, 2026 at 2:02 pm
Recently a loved one died, and I was asking close relatives for their opinion on grave design and colour. Many of them came back with a “this is what AI says” reply. It annoyed me so much. I want YOUR opinion based on how you knew her, FFS
@maybethatstoomanybooks
September 28, 2026 at 2:33 pm
wow, way to add salt to a wound
(p.s sorry for your loss, life sucks sometimes)
@TopdawgShadic
September 28, 2026 at 2:45 pm
It’s better to learn how to think, not what to think.
@THÚG々MOGAMBO
September 28, 2026 at 4:34 pm
@TopdawgShadic ETA
@AcselMitchell
September 28, 2026 at 3:03 pm
AI laughs?
@blackninja937
September 28, 2026 at 3:29 pm
Reject AI, embrace humanity
@Buff_Cupcake
September 28, 2026 at 3:34 pm
On one hand a massive chunk of our population will become idiots. But on the other hand, it will be incredibly easy to stand out just by doing the bare minimum.
@DKRSpeedline
September 28, 2026 at 3:58 pm
The question is: Why is ChatGPT choosing your ice cream? I don’t use AI anymore, and I am feeling great. Try to make your own decisions. You won’t regret!
@Squintillions
September 28, 2026 at 4:23 pm
Ok, so takeaway is that I’ve been missing out on the pralines and cream at Baskin-Robbins.
@weezypeezy1725
September 28, 2026 at 4:33 pm
AI is an exceptional tool to gather research tho, yes it could be making poor choices of the research it presents to you (and not presenting others), but can gather much clearer insights. A tool, with us as the operator and judge, not picking and just confirming what we want to hear. However, it seems like for most people that’s what it does— gives us just what we like, our favorites, like she was saying
@mawkernewek
September 28, 2026 at 5:39 pm
I’m a bit skeptical about AI’s ability to be incentivized to be more explorational, rather than more exploitation, since the regression to the mean effect based on being trained on vast datasets and creating a smooshed up average of them, is surely inherent to the way an LLM based AI works. I remember a few years ago, the Youtube algorithm had a bad reputation for sending people down rabbit holes by recommending increasingly extreme videos (e.g. alt-right, conspiracy theories) and there seemed to be an effort to sometimes show users something contrary.
@Pilgrim182
September 28, 2026 at 8:35 pm
Great message. Never thought about the idea of AI making humans so flat; like indeed robots, or animals. Insightful!
She didn’t use any slideshow or images, and I think that’s intentional, for the audience to feel the human essence at its highest. Thanks!!
One way we could escape this trap is reading more books, the old ones probably, and keep meeting with other people; friends, relatives, siblings, parents. Specially old people, they are the last ones from a generation without the internet
@Iceblockjoe
September 28, 2026 at 8:51 pm
YAHSHUA made me GOD of Salvation.
@angelavalentin
September 28, 2026 at 9:10 pm
I use AI, but for the last word I ask God.