TED talks are a great place for AI insights and inspiration, as they feature people with a range of AI backgrounds and experiences.
A search for “artificial intelligence” on the TED talks website turns up hundreds of results. In case you don’t have the time to watch them all, we compiled a list of our favorites.
Speaker: Anthony Goldbloom
Run time: 4:37
Why we like it: In 2013, a study out of Oxford University predicted that almost one in every two jobs had a high risk of being automated. Unfortunately, many people believe that their livelihood will be stolen by robots and machines—Goldbloom is here to remind us that isn’t necessarily the case. After all, machines excel at the repeatable, but they can’t outperform humans in novel situations.
Speaker: Sebastian Thrun and Chris Anderson
Run time: 24:22
Why we like it: Related to Goldbloom’s talk, Thrun reiterates that machines excel at repetitive work and we shouldn’t be afraid of AI. Rather, we should embrace the opportunity to explore more creative outlets while machines take on the more tedious tasks. This interview-style discussion is rich with interesting examples and firsthand experiences, making Thrun’s vision palpable.
Speaker: Tim Leberecht
Run time: 11:45
Why we like it: AI isn’t exactly known for being empathetic—As Leberecht puts it, machines “have no appreciation whatsoever for the unnecessary, the intimate, the incomplete and definitely not for the ugly.” But for our society to remain quintessentially human in the age of AI, these imperfections are necessary. For businesses embracing AI and machine learning, reminders like this are key to remaining authentic.
Speaker: Margaret Mitchell
Run time: 9:57
Why we like it: This talk is all about the weaknesses within AI and how we can help computers evolve to eliminate harmful biases and blind spots. Mitchell gives a poignant example of showing a machine learning program images of a house fire, to which it replied, “This is an amazing view! This is spectacular!” This particular blind spot arose because the computer had never been trained on high-contrast, red and orange photos of negative events, only positive. This anecdote and more paint a picture of how we can help AI evolve to create a better outcome.
Speaker: Tom Gruber
Run time: 9:47
Why we like it: As the co-creator of Siri, Gruber pulls from decades of experience with AI. In this talk, he explores how the AI of the future could impact things like humans’ ability for memory recall, bettering everything from interpersonal relationships to the detrimental effects of dementia. Watch along to see his vision for smarter machines and, subsequently, smarter humans.
Speaker: Oscar Schwartz
Run time: 10:57
Why we like it: Schwartz has a strong background in writing about the intersection of technology and human culture, and he draws upon that in this exploration of why humans react so strongly to the notion of a computer writing poetry. He explains that machines reflect back to us what we give to them—so we must pause and think about what idea of “the human” we want machines to reflect back at us.
Speaker: Zeynep Tufecki
Run time: 17:43
Why we like it: While a more cautionary tone than what we typically read and watch, this is a fascinating look at the ways AI fails, and what it means for the humans operating it. AI doesn’t follow human error patterns; so it’s hard to understand exactly what the machines are leaving out when they make decisions. For these reasons, it’s more important than ever to “ own up to our moral responsibility to judgment, and use algorithms within that framework,” as Tufecki puts it.
Speaker: Zeynep Tufecki
Run time: 22:56
Why we like it: Tufecki tackles a difficult topic head-on in this incredibly important talk. While we’re used to being targeted by online ads, we may not be aware of just how deep AI algorithms can go, if those in charge of them don’t act ethically. These types of challenges impact our broader world and it’s vital that AI trailblazers understand the risks.
In her talks, Zeynup Tufecki tackles an interesting and ever-evolving issue: The ethics, morals, and challenges we'll face while implementing AI. Beyond simply the adoption, how will AI impact the broader world?