Alphabet, the parent company of Google, has recently decided to sever its ties with Australian AI firm Appen, a key player in training various AI models, including Google’s chatbot Bard and search engine results. This decision, communicated to Appen over the weekend, will take effect on March 19, as revealed in a filing by Appen. Surprisingly, the company had no prior knowledge of Google’s intention to terminate the contract.
The termination of this partnership holds significant consequences, as Alphabet has been a major contributor, accounting for approximately one-third of Appen’s revenue. This move will notably impact around two thousand subcontracted workers affiliated with Alphabet, according to a statement by the Alphabet Workers Union released on Monday.
Appen, headquartered in Australia, has been a prominent player in the AI training field, boasting an impressive clientele that includes Microsoft, Apple, Meta, Google, and Amazon. Over 80% of Appen’s revenue has historically been generated from these five tech giants. With a global platform comprising about 1 million freelance workers across 170 countries, Appen has been a pivotal force in the AI industry.
In the fiscal year 2023, revenue from Appen’s collaboration with Alphabet amounted to $82.8 million out of its $273 million in total sales, according to the latest filing on Monday. However, despite its illustrious client base and nearly three decades of industry experience, Appen has faced challenges in recent years, marked by a loss of customers, executive departures, and financial setbacks.
In 2023, the company experienced a 30% drop in revenue, following a 13% decline the previous year. Appen attributes this downturn in part to “challenging external operating and macro conditions,” even as the demand for training data increased with the rise of generative AI tools.
Former employees, who preferred to remain anonymous, disclosed to CNBC in September that Appen’s current struggles to adapt to generative AI stem from years of weak quality controls and an organizational structure that lacks cohesion.
While Appen has undertaken various projects for tech companies in the past, such as evaluating search result relevance and assisting AI assistants in understanding diverse accents, the landscape has evolved. Presently, companies are investing more in processors from Nvidia and less in services provided by Appen, as large language models like OpenAI’s ChatGPT and Google’s Bard redefine the industry.
The history between Google and Appen includes previous conflicts, notably a wage dispute in 2019. Google mandated a minimum wage of $15 per hour for its contractors, a requirement that Appen allegedly did not meet, according to public letters from some workers.
Despite raises going into effect for Appen freelancers working on the Bard chatbot and other Google products in January 2023, labor issues persisted. In June of the same year, Appen faced charges from the U.S. National Labor Relations Board after allegedly firing six freelancers who publicly voiced their frustrations with workplace conditions. These workers were later reinstated.
Appen, in its filing on Monday, expressed its commitment to managing costs, revitalizing its business, and ensuring the delivery of quality AI data to its customers. The company intends to adjust its strategic priorities in response to the termination of the Google contract and promises to provide more details in its full-year results for FY23, scheduled for release on February 27, 2024.
What it means for freelancers
Freelancers in the AI training field, especially those associated with firms like Appen, should consider diversifying their client base. Relying heavily on a single major client, as seen in the case of Appen’s significant dependence on Alphabet, can leave freelancers vulnerable to sudden contract terminations, impacting their income and job stability.
Also, keeping abreast of industry dynamics, including shifts in demand for services and emerging technologies, is crucial. Appen’s struggles highlight the importance of adapting to changes in the AI landscape, particularly with the growing influence of large language models like ChatGPT and Google’s Bard.