Hello and welcome to Eye on AI. In today’s edition…What teens are saying about AI; Perplexity starts experimenting with ads; Greg Brockman returns to OpenAI; and a Sotheby’s AI art auction blows past expectations.
As AI rapidly changes industries, behaviors, and how society functions, adults can move forward having known the world before and after AI. Teenagers, on the other hand, are staring down an adulthood they know will look nothing like that of generations before. As they form relationships, develop their sense of self, prepare to look for work, and navigate an internet and media landscape shaken by AI, they will be particularly impacted by the decisions tech companies and lawmakers make—or don’t make—about AI today.
The Center for Youth, a youth-led research organization associated with the nonprofit Project Liberty, has dubbed today’s teenagers “Generation AI.” Led by two teens, the center this week published results from a survey of over 1,000 U.S. Teens about their usage, opinions, and fears of AI, adding to a growing body of research on the impact of AI on young people. The findings are an interesting look into how they’re using AI today and their fears for how AI will affect them tomorrow.
Around half of teens are using AI regularly
According to the survey, 47% of teens are using AI tools like ChatGPT several times a week or more. It doesn’t go into what they’re using AI for, but other reports have shed some light on this. One from nonprofit Common Sense Media—which found similar usage rates—says that teens are primarily using chatbots and AI search engines over image and video generating tools, leaning on them for homework, staving off boredom, and translation. Another report published by Hopelab and Harvard that focused on young people ages 14 through 22 similarly describes how they’re using AI for schoolwork, entertainment, companionship, and guidance—especially when it comes to questions they view as embarrassing or wouldn’t want to ask adults. It warns that “as generative AI use becomes more ubiquitous, adults should know that it may become the place teens go first.”
The Hopelab survey covers a slightly larger age range and cites a much lower rate of AI usage (only 15% use AI tools weekly or more, it says). Yet, the warning about AI being the first place teens may go hits hard in light of the death of Sewell Setzer III, a 14-year-old from Florida who killed himself after becoming increasingly obsessed with a Character.ai chatbot and relying on it for emotional support and guidance.
From self esteem issues to sextortion scams, society is still reeling from how social media has impacted the first generation of teens that grew up with platforms like Instagram and Snapchat, which dominated youth digital and social experiences without regulation or proper safeguards. All these surveys may feel redundant, but as we learned from the social media era, these are the types of impacts that need to be understood sooner rather than later.
Teens want regulation, not an AI takeover
The vast majority of teens view AI risks as a top issue for government regulation. According to The Center for Youth and AI survey, 80% said AI risks are important for lawmakers to address, ranking higher than social inequality (78%) and climate change (77%). Only healthcare access and affordability ranked higher, both at 87%.
Specifically, they’re worried about misinformation, deep fakes, mass surveillance, privacy violations, and AI taking over—throughlines that emerged in the Hopelab survey as well. Quotes shared from survey respondents in the Center of Youth AI report show teens expressings concerns that they never know if what they see online is real or AI-generated, that there will be no jobs available for them to work, and that we’ll lose what makes us human.
“I just hope that as AI gets more powerful, we don’t lose touch with what makes us human. I don’t want to live in a world where everything is just automated and we’re not needed anymore,” said one 17-year-old respondent.
And with that, here’s more AI news.
Sage Lazzaro
[email protected]
sagelazzaro.com
AI IN THE NEWS
OpenAI, Google, and Anthropic are hitting a wall in developing more advanced general AI models. Following reporting from The Information that OpenAI’s upcoming Orion model failed to surpass the capabilities of GPT-4 on some tasks, new reporting shows it’s not the only firm hitting a wall. The latest models being developed inside Google and Anthropic are also falling short of expectations and failing to provide the same leaps forward seen between previous model generations, Bloomberg and The Information reported. Timelines for releases are being pushed, raising doubts about the massive investments being made into AI. The firms are looking for new approaches as the “bigger is better” approach seemingly comes to an end.
Perplexity will begin experimenting with ads on its platform this week. The ads will be formatted as “sponsored follow-up questions” and will be generated by AI, not written by the brands. The ads will initially roll out to U.S. Users with Indeed, Whole Foods, Universal McCann, and PMG among the first advertisers. You can read more in TechCrunch.
OpenAI president Greg Brockman returns from leave of absence. Brockman stepped away in August, raising concerns he might not return and would be yet another executive to flee from the company this year. He shared on X that he’s back, and in an internal memo told staff he’s working with Sam Altman to create a new role in which he’ll focus on significant technical challenges, according to Bloomberg.
The EU begins a consultation on definitions of AI and unacceptable risks. The European Union’s new AI Office announced that it was launching a multi-stakeholder consultation on how the definition of AI in the EU AI Act may need to change in the future. It is also calling for stakeholders to provide examples of AI applications and uses that might pose an unacceptable risk.
FORTUNE ON AI
Elon Musk’s xAI safety whisperer just became an advisor to Scale AI —by Sharon Goldman
Europe’s AI industry watches Trump’s return with a mix of fear and hope —by David Meyer
Exclusive: Tessl worth a reported $750 million after latest $100 million funding to help it build ‘AI native’ software development platform —by Jeremy Kahn
AT&T’s CEO says AI may cause power shortages and it could be ‘the next big social issue in the United States’ —by Orianna Rosa Royle
Glassdoor CEO talks about the hottest jobs in the AI boom—and the one job he thinks is phasing out —by Emma Burleigh
This United Nations AI official explains why she doesn’t want an international agency for AI —by Emma Burleigh
AI CALENDAR
Nov. 19-22: Microsoft Ignite, Chicago
Nov. 20: Cerebral Valley AI Summit, San Francisco
Nov. 21-22: Global AI Safety Summit, San Francisco
Dec. 2-6: AWS re:Invent, Las Vegas
Dec. 8-12: Neural Information Processing Systems (Neurips) 2024, Vancouver, British Columbia
Dec. 9-10: Coins2Day Brainstorm AI, San Francisco (register here)
Jan. 7-10: CES, Las Vegas
EYE ON AI NUMBERS
$1.08 million
That’s how much an AI-created portrait of AI pioneer Alan Turing sold for in a Sotheby’s auction last week. The auction house had estimated it would go for between $120,000 and $180,000.
It’s not the first such sale of AI-created art but was a first for Sotheyby’s and unique in that, unlike most AI art that’s generated digitally by text-to-images models, this piece was also painted on canvas by an AI robot. I previewed the auction in the newsletter a few weeks ago, discussing what it means for debates around whether AI can be credited as an artist and the larger, increasing criticisms of the practice by human artists, and how software companies are trying to cash in.
