Today we’re back for part two of our roundup to close out the summer, looking at AI’s impact on education three years post-ChatGPT, and AI in the hiring process, from massive paydays to job shifts to the way people are interviewed and onboarded.
But first, in a demonstration of how fast AI is changing, we need to update a few of our stories from last week.
In our coverage of the surge in AI investments, we talked about the $10 billion committed to Anthropic, raising its value to $170 billion in August. Not long after rolling out our article they secured another $13 billion, meaning they’re now valued at $183 billion.
And Nvidia, who led our last piece with its incredible gains over the past decade, has been on a continued slide in the markets. Note that they’re still (as of the time of this writing) easily the world’s most valuable company (at $4.1 trillion).
Back-to-School AI Update
Teachers and students are all coming back to school at various levels with increasingly more experience with AI, and more accessible and capable tools at their disposal.
It’s been three school years since chatbots became widely available, and one thing’s for sure—students are wholly on board with AI use at almost every level.
K–12: AI in the Classroom
While teachers may be saving themselves six hours a week through AI use, most agree schools and educators are trailing students in use at almost every level.
A Common Sense Media/Ipsos survey from 2024 showed that the top AI use case for teens (at 53%) was help with homework.
In the interest of helping teachers catch up (and getting their products used in schools), major AI providers are making partnerships both to help train educators and provide tools for the classroom.
One of these is the National Academy for AI, which partners companies like Microsoft, OpenAI, and Anthropic with the American Federation of Teachers to support 400,000 of the union members over the next five years, or around 10% of all teachers nationwide. The $23 million initiative will include tools, workshops, and online courses, and count for continuing education credits.
One new tool built around educational use and released in this period is OpenAI study mode. Added to ChatGPT with the goal of discouraging homework shortcuts, it acts like a Socratic tutor—asking questions and prompting reasoning before answers. It’s available on the free tier and also being evaluated by Stanford learning experts, though critics note students can still just toggle back to normal ChatGPT.
AI in Education: A University Perspective
“They gotta get dynamic or die.”
This is a quote from Princeton historian D. Graham Burnett on the necessity of American institutions to face the ongoing transformation of AI in education head-on. He appeared on the Hard Fork podcast last Friday and wrote a recent piece for the New Yorker titled Will the Humanities Survive Artificial Intelligence?
College and university students worldwide are heavily using AI, though right now it may be causing what Burnett called a crisis for the “police function” of educators—making assignments like traditional papers useless.
But one assignment he gave students—to have a conversation with an AI chatbot on a provided topic, edit that to four pages and turn it in—he found to be an incredible demonstration of the power AI has to help students learn.
And while many students are demonstrating highly effective means of using AI for filling knowledge gaps, synthesizing lecture notes, and generating quizzes and self-checks, many are also using it to cut corners and bypass work meant to reinforce their education.
As the technology continues to improve, the question remains not only how AI will help students learn in the near term, but also how it changes what they will need to learn and why.
The AI Impact on Hiring and the Jobs Market
From school to the newly (and not-so-newly) graduated.
The New York Times in August ran a pair of articles about the change in Silicon Valley’s tech culture, charting a move from a perks-heavy workplace to one marked more with layoffs, in-person hubs, and even RTO.
This comes as companies like Amazon, after warning of AI-triggered cuts in June, now has divisions that aim to first promote staff with AI skills, or who can detail “innovative thinking” through AI. Their smart-home businesses (RBKS) want to see employees and managers alike able to document how they’re doing “more with less.”
Google, too, has had internal meetings about its employees feeling “the urgency” and becoming “more AI-savvy” and productive in day-to-day tasks. The goal, as at Amazon, is to increase productivity with AI, and the company is promoting internal training and tools provided for this purpose.
Yet this sense of tightening and pressure comes as those at the top end of AI are seeing higher salaries than ever before.
Blitz hiring, where companies buy leadership or top talent without the company itself, is seeing increased use alongside the record surge in mergers and acquisitions.
About That Meta AI Talent Hiring Spree
After losing top talent and delaying a new Llama rollout, Meta went on a well-publicized hiring spree in their bid to avoid falling behind in AI.
Those efforts were detailed last time out with signing bonuses and salaries reportedly reaching nine figures in extreme cases.
And during this period, more details emerged, like the targeting of 12 of 50 employees at the well-funded Thinking Machines Lab, with one of the offers worth over $1 billion over multiple years. (Per Wired, most were in the $200–500 million range over four years.)
And late last week (September 4) Bloomberg reported on another big hire, with Apple’s leading robotics researcher being the latest high-profile talent to jump ship for Meta.
All this poached talent from other big AI companies is now going to work at Meta, with news about reorganizations and adjustments, and at least three resigning in just weeks (two going back to OpenAI).
The Superintelligence/TBD effort is now part of an AI operation organized into four groups (one on superintelligence, one research, one products, and one infrastructure).
And while CEO Mark Zuckerberg talked to investors about developing AI that can self-improve and spending “hundreds of billions” on data centers to support the effort, work is also underway within the Superintelligence team to get a new Llama model released before the end of the year.
AI and Entry-Level Jobs: Goldman Sachs and the Stanford AI Jobs Study
Last time out we talked about data indicating a freeze in some entry-level hiring being attributed to AI. This was further supported in August by research from Goldman Sachs and Stanford University.
Joseph Briggs, a senior global economist of Goldman’s research division, told CNBC that most companies haven’t deployed AI in production cases, which is limiting overall job impact. But younger workers in tech, whose jobs are the easiest to automate, are giving the first solid signs of a displacement trend.
Companies are now holding off on hiring inexperienced new talent, but Briggs said he believes 6–7% of all workers could lose their jobs from AI automation, and that’s without factoring in AGI.
A new research paper from the Stanford University HAI finds a more complicated picture. While their research of ADP data from 2022–2025 provides solid evidence that jobs are being eliminated for inexperienced workers (up to a 13% decline for workers aged 22–25 in areas like customer service and software development), it also shows flat or increasing opportunities for workers with experience. It also shows that wages are not being reduced.
And with their data pointing to so-called “centaur” partnerships of AI and humans outperforming what AI systems can do on their own, the researchers argue AI needs human partnership, and advocate for more research in this area.
The Future of AI Recruiting and Impact on the Workforce
AI is already transforming hiring.
With numerous systems in play that enable interviewees to cheat and get away with it, a number of companies (including Google, Cisco, and McKinsey) are once again requiring face-to-face interviews.
But Meta may be going the opposite direction. Reports from 404 Media describe how the company is working to test coding interviews that allow AI assistants to be used in their coding interview process.
Given a desire to encourage AI assistant use at work, “it should be no surprise” they’re working on this option, a company spokesperson said.
And speaking of the recruiting process, a surge in AI-generated application assistance by candidates is accompanying the rise in use of AI in hiring, for matching and pre-screening candidates (80% of companies use AI to review resumes), contacting candidates (40% here, per a Resume Builder survey), and even running actual screenings.
PTP is among these companies, working with the centaur goal of pairing our experienced and talented recruiters with the best solutions in AI. The goal for us is to reduce time-to-hire and costs while maintaining quality and affording the broadest range of options for our clients.
And research profiled by Bloomberg from University of Chicago and Erasmus University Rotterdam has found successes here for businesses, with interviews by AI agents 12% more likely to lead to job offers, with workers hired in this process also 17% more likely to stay on the job for more than a month.
Ironically, the study found the bots spoke less and listened more and were less impacted by repetition and workload.
On the flip side, Gartner found that only a quarter of candidates feel AI evaluates them fairly, and that six in 10 would be more likely to apply if in-person interviews were offered.
Yahoo Finance’s Senior Columnist Kerry Hannon offered these tips on AI interviews for job seekers:
- Treat the AI like a human, because it’s looking for the same things in you
- Consider it an audition, like making a recording, since even though you may not be talking to another person, your performance will still be evaluated
- Keep it as professional as with humans and check your equipment beforehand
- Stay focused because unlike humans, the AI can always be looking at you
- Expect behavioral questions, as with humans
- And, as with humans, ask your own questions, stick to the facts, stay on target, and say thank you when it’s over
In short, AI recruitment tools may be machines, but they’re looking for the same things that human recruiters look for, and the biggest challenge could be approaching it in the same way you would any other interview.
Conclusion
That concludes our roundup for this period. Due to the increase in AI news, from here we’ll be going to a single, monthly roundup (vs bi-monthly), with our next (for September), coming at the end of this month.
Check out Part 1 in case you missed our last roundup, and if you need to catch up on AI overall, you can find our last six below (with the full run on PTP):
References
Generative AI in Higher Education: Evidence from an Elite College, arXiv:2508.00717 [econ. GN]
A.I. School Is in Session: Two Takes on the Future of Education, Hard Fork Podcast
Three in 10 Teachers Use AI Weekly, Saving Six Weeks a Year, Gallup
The Dawn of the AI Era: Teens, Parents, and the Adoption of Generative AI at Home and School, Common Sense Media
Google execs say employees have to ‘be more AI-savvy’ as competition ramps up and AI is already impacting the labor market, starting with young tech workers, Goldman economist says, CNBC
AI Is Forcing the Return of the In-Person Job Interview, The Wall Street Journal
If Amazon Ring staff want a promotion, they must now prove they’ve used AI, Fortune
Anthropic Completes New Funding Round at $183 Billion Value and Study of 67,000 Job Interviews Finds AI Outperforms Human Recruiters, Bloomberg
Your next job interview might be with an AI bot. Here’s how to ace it., Yahoo Finance
Meta is racing the clock to launch its newest Llama AI model this year, Business Insider
Researchers Are Already Leaving Meta’s New Superintelligence Lab, Meta’s AI Recruiting Campaign Finds a New Target, and Microsoft, OpenAI, and a US Teachers’ Union Are Hatching a Plan to ‘Bring AI Into the Classroom’, Wired
Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence, Stanford University HAI
Meta Is Going to Let Job Candidates Use AI During Coding Tests, 404 Media
