SAN FRANCISCO -- Luis von Ahn hoped to send a clear message to his 900 employees at Duolingo: Artificial intelligence is now a priority at the language-learning app.

The company would stop using contractors for work AI could handle. It'll seek AI skills in hiring. AI would be part of performance reviews, and it'll only hire people when things can't be automated. The details, outlined in a memo in April and posted on professional networking site LinkedIn, drew outrage.

Some cringed at AI translations, suggesting that learning languages needs human context. Many users threatened to quit Duolingo. Others blasted the company for choosing AI over its workers. The backlash got so loud that three weeks later, von Ahn posted an update.

"To be clear: I do not see AI as replacing what our employees do (we are in fact continuing to hire at the same speed as before)," von Ahn wrote in the update on LinkedIn. "I see it as a tool to accelerate what we do, at the same or better level of quality. And the sooner we learn how to use it, and use it responsibly, the better off we will be in the long run."

From Duolingo to Meta to e-commerce firm Shopify and cloud storage company Box, more companies are mandating their executives and teams implement AI-first strategies in areas such as risk assessment, hiring and performance reviews. Some of the directives are being detailed in public memos from top leaders, in some cases spurring outrage. Others are happening behind closed doors, according to people in the tech industry. The implication: AI is increasingly becoming a requirement in the workplace and no longer just an option.

"As AI becomes more popular and companies invest more heavily ... the tools will start being embedded in the work (people) already do," said Emily Rose McRae, an analyst at market research and advisory firm Gartner. "The work they do will change."

Meta plans to replace humans with AI to assess risk as it relates to the privacy reviews of its products, according to NPR. The company told The Post it's rolling out automation for "low-risk decisions," like data deletion and retention, to allow teams to focus on more complicated decisions.

At Shopify, everyone is expected to learn how to apply AI to their jobs, CEO Tobi Lütke wrote in an April 7 memo to staff. The tech would also be a part of prototyping, performance and peer review questionnaires, and teams would have to demonstrate why AI couldn't do the job before requesting new hires.

"AI will totally change Shopify, our work, and the rest of our lives," said Lütke's memo, which went viral on social media. "We're all in on this!"

In response, some people pushed back. "I cannot in good conscience recommend any company that is so recklessly throwing their good humans to the wind while putting all their faith in computer code that does not work a good portion of the time," Kristine Schachinger, a digital marketing consultant, responded to Lütke on X.

Following von Ahn's memo, Duolingo Chief Engineering Officer Natalie Glance shared details about what the strategy meant for her team. AI should be the default for solving problems and productivity expectations would rise as AI handles more work, she wrote. She also advised her team to spend 10% of their time on experimenting and learning more about the AI tools, try using AI for every task first, and share learnings.

"We don't have all the answers yet -- and that's okay," she wrote in an internal Slack message she later posted on LinkedIn.

"Duolingo was not the first company to do this and I doubt it will be the last," said Sam Dalsimer, Duolingo spokesman.

Many white-collar workers across industries are already using AI in their work, from summarizing documents to writing emails and reports to research and data crunching. And Microsoft, Zoom and Google have been increasingly packing AI into many of their work products.

At Box, CEO and co-founder Aaron Levie expects the company to use AI to eliminate tasks that clog up people's workflows, automate more so that the company can reinvest savings to achieve more and foster more experimentation. He also wants to train employees to be proficient in AI.

"In engineering, we're probably in the final generation where you can go into a company with no AI coding expertise," Levie said in an interview with The Post. "What we've found internally and with more and more customers is that AI is really helping accelerate doing more, getting better work done and solving problems faster."

AI is already becoming a factor in some job interviews. Google software engineering candidates are given access to AI tools to solve problems, said Ryan J. Salva, senior director of product at Google. While Google doesn't have any mandates on using AI, it expects people to get the job done in the most efficient way, including using AI.

"I would be surprised if they did not use AI most of the time," he said about developer interviews.

Some directives are less about rules, and more about pushing employees to "wake up" and use AI. "Here's the unpleasant truth," Micha Kaufman, Fiverr founder and CEO, wrote in a memo. "AI is coming for your jobs. Heck it's coming for my job too."

His memo advised employees to become prompt engineers and master the latest AI solutions in their fields. He then spent three hours huddled in a meeting room with 250 workers discussing his expectations, including that new hires be proficient in AI, employees double or triple their output and strive to automate their jobs. The goal is to redirect people to tasks that only they can do, such as generating innovative ideas or using human judgment, Kaufman said in an interview.

"If you don't do this now, your value in the industry is going to drop, and you're going to be doomed," he said.

Similarly, Wade Foster, co-founder and CEO of workflow automation platform Zapier, posted on X that he now expects aspiring employees to be equipped with AI skills. "We're setting a new standard at Zapier. 100% of new hires must be fluent in AI."

This "AI-first" mandate isn't being well-received by everyone. People worry about AI replacing them at work, lowering the quality of products and services, and creating problematic outcomes such as misinformation and errors.

"AI first = Employees last," said Chris Craig, a West Hollywood founder and CEO. "This is poor leadership."

Some companies, lured by the promise of efficiency, are expected to adopt AI without making necessary workflow and structural changes, McRae of Gartner said. As a result, they'll probably fail to get their return on investment and frustrate workers in the meantime, she added.

Moving too fast could backfire. After heavily touting its AI focus, Klarna, a Swedish fintech company, recently pulled back.

AI helped the company lower vendor expenses, increase internal productivity and better manage operations with fewer employees, according to a company filing submitted with the Securities and Exchange Commission in March. From 2022 to 2024, Klarna shrank its headcount by 38% (Klarna said its AI assistant handles tasks equivalent to more than 800 full-time roles). But the company ultimately went too far in its attempts to cut costs, Klarna CEO Sebastian Siemiatkowski told Bloomberg in May. The company told The Washington Post it recognizes AI's limitations, especially in nuanced interactions, and plans to recruit more human gig workers while investing in the tech.

"We've learned that successful deployment hinges on thoughtful implementation," said Clare Nordstrom, a Klarna spokesperson. "Quality customer experience requires balance, which is why we're investing in advanced technology with a human touch."

At many startups, AI-first strategies are nothing new, said Roy Bahat, head of venture capital firm Bloomberg Beta, which invests in the future of work and AI startups.

"Most startups don't even need to think about having a policy about (AI) because if they're not already built from the ground up using AI, they're missing something."

AI will steadily make its way into some workplaces regardless of mandates, McRae said. That's because increasingly, generative AI will be integrated into workers' tools seamlessly. They may use AI and not even realize it, she said. But until then, she expects more companies to join the AI-first bandwagon.

"Everyone needs to understand that the tools we're using are new and our understanding of them will change."