Welcome to Eye on AI! I’m Sharon Goldman, AI reporter at Fortune. I’m excited to share that I’ll be taking the helm on the Thursday edition going forward. Don’t hesitate to reach out and let me know what you think! In today’s edition: Musk’s DOGE uses AI to snoop on federal workers…a Meta exec denies the company artificially boosted Llama 4’s benchmark scores…Europe unveils plan to become ‘AI continent’ with simpler rules, more infrastructure…Google uses AI to recreate The Wizard of Oz for the Las Vegas Sphere

On Monday, Shopify CEO Tobias Lutke shared on X an internal memo he sent to his over 8,000 employees. The memo, which mostly focused on requiring Shopify employees to use AI effectively and often, went viral and set off a firestorm because of one spicy sentence: “Before asking for more headcount and resources, teams must demonstrate why they cannot get what they want done using AI.”

Maybe you don’t think today’s AI models and tools will replace knowledge-work jobs imminently, or lead to mass layoffs in the short-to-medium term. Maybe you see generative AI simply as the smart, reliable assistant you always wanted. But I think Shopify is saying the quiet part out loud: We are headed into an AI-first era where new human hires will soon be required to prove their worth among a sea of digital “employees.” There is no doubt in my mind that AI is already starting to replace new hiring in knowledge industries—but most CEOs will not yet admit it publicly.

I worry about that silence—a lot. I worry that CEOs, by keeping quiet, are not preparing their employees, and the public, for what’s quickly coming down the pike. I’m concerned that workers don’t use AI tools enough yet, in the right ways, to understand how their roles, and their department, and their companies, will change. I’m troubled that our society is not yet ready to support the consequences of these changes with the right communication, upskilling, and, yes, financial and mental health support.

Employees aren’t dumb. They know AI is coming.

Yet I also know that workers are not stupid. We all watched Elon Musk’s DOGE lay off thousands of federal workers in the name of efficiency, while making it clear the U.S. government wants to use AI to boost productivity and reduce the size of the federal workforce. Most of us are not so arrogant that we can’t see that even if an AI tool cannot do everything we do, and makes mistakes, a company may still decide it does enough to eliminate the need for that junior team member.

CEOs, after all, are under pressure—from board members, investors, and shareholders—to deliver on growth and revenue. Particularly in the tech industry, CEOs may soon have little choice but to deliver a lean, mean, AI-powered machine.

Yesterday, I spoke to Micha Kaufman, CEO of Fiverr, an online marketplace for freelance services with around 1,000 employees. He said he issued his own internal memo as a wake-up call to employees, telling them to boost efficiency through AI rather than expecting new hires and an expanded team. His language, while slightly less brusque than Shopify, was essentially the same: “It does not make sense to hire more people before we learn how to do more with what we have,” said Kaufman in his message, adding that he was sharing an unpleasant truth: “AI is coming for your jobs,” the message read. “Heck, it’s coming for my job too. This is a wake-up call.”

Radical Candor

The idea for the message had been “building up for quite a while,” said Kaufman. He had given a talk about AI the day before and had met with many CEOs. “There was a sense that this is becoming the consensus,” he said. “I do think there is going to be a wave of layoffs in tech, and also an expectation of better performance, better output from existing team members.”

He said he felt the need to offer “radical candor” to his employees, which he said means calling the reality what it is “because you care about people—not because you don’t care, or because you want to manipulate people. You want to wake them up.”

I’m not sure how already stressed and overworked employees facing an uncertain economy feel about Lutke and Kaufman’s “radical candor.” But as a daily user of ChatGPT and other AI tools, I appreciate the fact that a few CEOs are telling it like it is. ChatGPT and its ilk has been a personal game-changer as big as my first laptop, Google search, Twitter, and the advent of cut-and-paste combined. It’s hard to describe how indispensable it has become for everything from simple searches and deep research to summaries, editing, and idea feedback. Writing this newsletter was sped up considerably, for example, thanks to the back and forth I had with ChatGPT to tweak my initial headline; quickly adjust my initial efforts to summarize existing news stories; act like a thesaurus on steroids; and uncover the report I highlighted in Eye on AI Numbers rather than searching for a needle in a haystack using Google search. But that kind of regular use and experimentation has also made it clear that we are just at the beginning of how these tools will impact a company’s human resources.

A recent Pew Research study conducted in October 2024 found that 55% of American workers reported that they rarely or never use AI chatbots like ChatGPT, Gemini, or Copilot in their jobs, and an additional 29% indicated they hadn’t heard of these tools at all. Perhaps that number has already changed in the past six months—and perhaps some of those workers are in jobs where generative AI would likely not be useful.

Still, on behalf of the knowledge workers out there, I’m here to say the future workplace will be far more complicated than AI “peacefully coexisting” with humans rather than replacing them, as Workday CEO Carl Eschenbach told me recently. Sure, AI may coexist with some human knowledge workers. But companies certainly won’t need as many of us as they currently do. We all need to speak out more about what happens next.

And with that, here’s the rest of the AI news.

Sharon Goldman
sharon.goldman@fortune.com
@sharongoldman

AI IN THE NEWS

Musk’s DOGE using AI to snoop on federal workers. Elon Musk’s Department of Government Efficiency (DOGE) is reportedly using AI to monitor U.S. federal employee communications, in at least one agency, for expressions of disloyalty toward President Donald Trump and his administration, according to Reuters. At the Environmental Protection Agency, managers were told by Trump appointees that Musk’s team is rolling out AI to monitor workers, including looking for language in communications considered hostile to Trump or Musk. After the story was published, the EPA acknowledged in a statement that it was “looking at AI to better optimize agency functions and administrative efficiencies” but said it was not using AI “as it makes personnel decisions in concert with DOGE.”

Meta exec denies the company artificially boosted Llama 4’s benchmark scores. A Meta executive denied a rumor the company had gamed performance benchmarks, according to TechCrunch. The executive, Ahmad Al-Dahle, VP of generative AI at Meta, said in a post on X that it’s “simply not true” that Meta trained its Llama 4 Maverick and Llama 4 Scout models on “test sets,” or collections of data used to evaluate the performance of a model after it’s been trained. Doing so could inflate a model’s benchmark scores and make the model appear more capable than it actually is.

Europe unveils plan to become ‘AI continent’ with simpler rules, more infrastructure. The European Union presented a plan on Wednesday to help its AI industry compete more aggressively with the U.S. and China, according to CNBC. In a press release, the European Commission, the executive body of the EU, outlined its so-called “AI Continent Action Plan,” which aims to “transform Europe’s strong traditional industries and its exceptional talent pool into powerful engines of AI innovation and acceleration.” The plan, which comes just a year after the EU AI Act was formally adopted, includes a commitment to build a network of “AI factories” and create specialized labs designed to improve the access of startups to high-quality training data. The EU will also create a new AI Act Service Desk to help European firms comply with the law.

Google used AI to recreate The Wizard of Oz for the Las Vegas Sphere. According to the Wall Street Journal, the Las Vegas Sphere’s 160,000-square-foot curved, immersive screen will show the 86-year-old classic film The Wizard of Oz. Google’s enterprise business, Google Cloud, and its research unit, Google DeepMind developed new AI methods to enhance resolution and extend backgrounds to include characters and scenery not in the original shots, techniques called “performance generation” and “outpainting.” Now, AI has touched over 90% of the movie, which will debut at the Sphere on August 28.

FORTUNE ON AI

Google’s latest AI model is missing a key safety report in apparent violation of promises made to the U.S. government and at international summits —by Beatrice Nolan and Jeremy Kahn

Google Cloud moves deeper into open source AI with Ai2 partnership —by Jeremy Kahn

Here’s why tariff shock could hit AI’s data center boom —by Sharon Goldman

Commentary: AI’s job boom? Not before the bust —by Nada R. Sanders and Paula Caligiuri

AI CALENDAR

April 9-11: Google Cloud Next, Las Vegas

April 24-28: International Conference on Learning Representations (ICLR), Singapore

May 6-7: Fortune Brainstorm AI London. Apply to attend here.

May 20-21: Google IO, Mountain View, Calif.

July 13-19: International Conference on Machine Learning (ICML), Vancouver

July 22-23: Fortune Brainstorm AI Singapore. Apply to attend here.

EYE ON AI NUMBERS

$4.8 trillion

That’s what the valuation of the global AI market will be by 2033, according to a UN Trade and Development (UNCTAD) report. The organization projects the current $189 billion global market will see a 25-fold increase in just a decade, fast becoming the “defining technology of our time.”

However, it emphasized that AI development is highly concentrated: Today, the US and China hold 60% of all AI patents and produce a third of global AI publications. In addition, today’s AI landscape is shaped by multinational tech giants like Apple, Nvidia, and Microsoft, which each have a market value of around $3 trillion—rivaling the GDP of the entire African continent.

Developing countries need to catch up in setting national AI strategies: By 2023, the report said that two thirds of developed economies had one, but only 30% of developing countries did (excluding the least developed). Among least developed countries, only 12% had a strategy in place. The report calls for “inclusive AI governance” and stronger international cooperation.

Shopify is saying the quiet part out loud: AI will replace new hiring—other CEOs just won’t admit it