Duolingo's Mass Layoffs Spark Concerns: Is AI Bringing a Wave of Unemployment?
-
There is no doubt that we are rapidly entering the era of generative AI. Since OpenAI launched ChatGPT over a year ago, the tech industry has been swept up in a generative AI frenzy, with giants like Google, Microsoft, and Meta quickly following suit, rolling out their own large models and applications: integrating generative AI features to deliver entirely new product experiences.
However, the impact of generative AI is not limited to internet companies and the tech industry. Like the PC and internet technologies before it, generative AI is ushering in a new revolution in productivity technology, with profound implications for production and organizational methods.
Compared to previous AI represented by smart assistants, generative AI like ChatGPT, based on deep learning and natural language processing capabilities, can better understand and generate human language, producing high-quality text through dialogue and completing tasks such as writing reports, replying to emails, creating images, and translating text.
A year after ChatGPT's release, generative AI itself is transitioning from cloud-only to a hybrid model of cloud and edge computing. Whether in smartphones or computers, the latest high-performance processors from chip giants support AI engine computing power of 75 TOPS, enabling the deployment of generative AI models on the edge.
However, while this impending AI technology revolution greatly enhances work efficiency and productivity, will it bring another direct consequence—mass unemployment? This has been a hot and concerning topic in technology and social discussions over the past year.
Just last month, Duolingo, the world's largest foreign language learning community, suddenly laid off thousands of contract translators, replacing their jobs with AI products. Even the remaining contract translators now only review AI-generated content to ensure translation quality.
Duolingo's decision to replace human translators with AI on a large scale has sparked considerable discussion on the American internet. Amid growing concerns about AI replacing human jobs, Duolingo's move undoubtedly validates many people's worst fears: AI tools are leading to mass unemployment.
It should be noted that Duolingo dismissed external contractors, not internal full-time employees. External contractors are a common practice in the tech industry to reduce hiring costs. Duolingo also employs external contractors worldwide for various languages to translate learning content and user interfaces for its foreign language learning software.
On the American online forum Reddit, a laid-off contractor named Nick Ciardi posted termination emails from Duolingo, confirming the large-scale layoffs. He stated that he received the termination notice in mid-December, and in his core team of four members, two were laid off while the remaining two will now focus on reviewing AI-generated content for quality assurance.
In Duolingo's Reddit community, many users expressed gratitude for the contributions of these contract translators while lamenting the inevitable replacement of human translators by AI. One user wrote, "Unfortunately, it's only a matter of time before AI replaces humans in the translation field."
However, some users argue that AI currently cannot completely replace high-quality human translation. Accurate translation requires not only fluency in both languages but also cultural understanding. Even Google Translate, currently the most advanced AI translation system, often produces significant errors despite conveying basic sentence meanings.
Additionally, some users complained about numerous unnatural expressions in Duolingo's German version, suggesting poor quality control in previous human translations. With AI taking over, there are concerns that Duolingo's translation quality might deteriorate further.
Over the past year, mass layoffs have become commonplace in Silicon Valley. Tech giants like Google, Meta, Microsoft, and Intel have all cut tens of thousands of employees, making Apple—which hasn't conducted layoffs—an outlier.
However, most tech companies have primarily laid off employees hired during their expansion phases in recent years. This adjustment stems from tightened capital markets, persistent global recession fears, and concerns over future business growth.
So why did Duolingo choose to terminate thousands of contract translators right before Christmas, leaving them in distress during the holiday season? Is it due to financial difficulties or bleak growth prospects?
Founded in 2011 at Carnegie Mellon University in Pittsburgh, Pennsylvania, Duolingo was created by two entrepreneurs from Guatemala and Switzerland. Computer science professor Luis von Ahn launched the company with the direct goal of helping his fellow Guatemalans learn English more affordably. He has served as Duolingo's CEO and chairman since its inception.
Over its more than ten years of development, Duolingo has received investments and support from Silicon Valley venture capitalists like Union Square and tech giants such as Google. Hollywood actor Ashton Kutcher was also an early investor in Duolingo. In June 2021, Duolingo went public through an initial public offering (IPO) and currently employs approximately 650 full-time staff. In addition to language learning, Duolingo offers online courses in subjects like math and music.
In fact, Duolingo has always been an enthusiastic advocate and user of AI. Founder and CEO Luis von Ahn told VentureBeat in an interview that AI technology is applied in nearly every aspect of the app, from recommending learning challenges to users to specific language training dialogues—AI helps determine users' learning progress.
AI has also brought Duolingo new service packages and revenue streams. Just last year, Duolingo launched Duolingo Max, a new GPT-4-powered subscription service priced at $30 per month or $168 annually. Subscribers to Max receive AI-generated explanations for each exercise, a feature not available in other subscription tiers.
From any perspective, Duolingo currently faces no performance pressure that would force layoffs. They are in the best period in the company's history, with both user growth and revenue surging rapidly. The stock price is also at its highest historical range, having increased by 187% over the past year, with a current market capitalization exceeding $8.7 billion.
During the layoff wave at the end of 2022, Duolingo publicly stated that they did not need to lay off employees nor freeze hiring. This confidence is backed by solid performance. In the 2023 fiscal year, Duolingo's revenue reached $484 million for the first time, with quarterly revenue growth exceeding 40% year-over-year. Moreover, the company successfully turned a profit starting last year.
In the most recent third quarter of last year, Duolingo's monthly active users grew by 47% year-over-year to 83.1 million, while daily active users increased by 63% to 24.2 million. Paid subscribers rose by 60% to 5.8 million. Revenue for the third quarter was $138 million, up 43% year-over-year, with a net profit of $2.81 million, a 115% increase compared to the previous year.
Duolingo isn't facing significant external pressure to cut costs, nor is it confronting a bleak outlook of slowing or shrinking business growth. Given such favorable circumstances, its large-scale termination of contract workers before Christmas can only be interpreted as a move by a publicly traded company pursuing financial performance. Since these weren't full-time employees being let go, it doesn't count as mass layoffs, so Duolingo didn't even issue a statement or need to respond publicly.
From the perspective of Blake Vente, a machine learning software engineer at Audible, this incident could harm everyone: (declining translation quality) would affect user experience, mass contract worker terminations would damage employee morale, and ultimately impact company culture. If users start viewing Duolingo as just an AI-wrapped product company, they might eventually switch to personalized generative AI products like MemBot instead of Duolingo's expensive service packages.
Although Duolingo's termination of contract workers is a normal business decision for a public company to reduce expenses, the pre-Christmas layoffs have still sparked considerable controversy. Capitalists are too ruthless. As AI gradually replaces routine jobs, after translation gets eliminated, will my position be next?
Following Duolingo, Google may be the next company to experience mass layoffs due to large-scale AI adoption. Reports indicate that Google is replacing traditional Google Ads with AI tools that possess more personalized and natural language capabilities. This means AI can search vast websites for advertisers, generate relevant keywords, slogans, descriptions, and images, greatly optimizing the ad service generation process.
While this clearly enhances the user experience of Google's ad services and improves efficiency, it also implies that Google no longer needs as many advertising sales employees. Consequently, Google is considering restructuring its advertising sales department, which could result in significant job losses among its current 30,000+ advertising sales staff.
Having just celebrated its 25th anniversary, Google is currently facing a mid-life crisis with slowing business growth and challenges to its core competitiveness. Google CEO Sundar Pichai has publicly stated the need to take proactive measures to avoid future setbacks. At the beginning of last year, Google already laid off 12,000 employees. If Pichai decides to implement major layoffs in the advertising sales department in 2024, it may not come as a surprise.
During last year's Hollywood strikes, generative AI tools were a key negotiation point between labor and management. Disney's use of AI to create opening animations has also sparked significant controversy. Writers, actors, and post-production staff worry that studios' extensive use of AI tools, trained on their existing works, could lead to forced pay cuts or even job losses in the future.
Their concerns are not unfounded. The film industry is increasingly adopting AI tools, with major studios like Warner Bros., Netflix, and Disney already using AI for tasks such as special effects, subtitle generation, and script summarization. Streaming giants like Amazon, Netflix, and Apple, with their strong technological capabilities and top-tier talent, further accelerate this trend. The full integration of AI into film production seems inevitable—it's only a matter of time.
Last summer, Disney's Marvel directly used AI to generate the opening sequence for the TV series Secret Invasion. Even traditionally risky animal performances, like those involving lions, were replaced with AI-generated imagery, eliminating the need for live animals on set.
The Writers Guild has finally reached an agreement with production studios. In addition to conditions such as increased compensation and secondary royalties, Hollywood producers have also compromised on AI tools, pledging not to use AI to write or rewrite literary content. AI-generated content cannot be considered source material, meaning it must not affect writers' creative rights or individual entitlements.
During the five-month-long strike, writers were willing to halt the industry and forgo their salaries to secure concessions from producers, achieving a temporary victory. However, in the next labor negotiations three years from now, writers may face even greater challenges as AI capabilities potentially undergo significant upgrades.
Will AI lead to mass unemployment? On this issue, pessimists and optimists hold entirely different views. However, no one can deny that the widespread adoption of generative AI tools will indeed replace many routine and procedural jobs, reducing the demand for such positions. The most direct example is that AI image-generation tools like Midjourney have already begun replacing traditional graphic designers and illustrators.
In a previous interview with Sina Technology, Hari Srinivasan, Vice President of Product Management for LinkedIn's Talent Solutions, stated that historically, technological advancements have always altered traditional job roles and responsibilities, leading to the disappearance of certain positions. However, he emphasized that such changes also drive innovation and creativity, generating new job opportunities. Srinivasan stressed the importance for professionals to acquire new skills and learning methods to adapt to these shifts.
Updating one's skill set to meet the demands of the AI era has become a critical factor in maintaining workplace competitiveness. According to LinkedIn's report, by 2030—within the next eight years—65% of global job skills will undergo transformation. Over the past eight years, 25% of job skills have already changed.
In other words, for professionals to remain competitive by 2030, they must update nearly two-thirds of their skill sets, focusing on areas where AI cannot yet replace human capabilities. Failure to do so may result in being unable to meet workplace demands in the AI era and facing the risk of obsolescence.