OpenAI has announced plans to nearly double its global workforce to 8,000 employees by the end of 2026, signaling that the race to lead artificial intelligence development has entered a new phase of aggressive scaling. The announcement comes alongside a parallel trend: thousands of ordinary people are now selling access to their voices, faces, and written communications to AI firms — a gig economy offshoot that is as controversial as it is lucrative.
OpenAI Hiring Boom
The hiring announcement marks one of the most aggressive workforce expansions in the technology sector this year. OpenAI, which employs approximately 4,500 people today, intends to grow to 8,000 by December 2026 — an increase of roughly 3,500 positions. Roles span engineering, research, safety, policy, and operations, with new hubs planned in Singapore, London, and Toronto.
Why OpenAI is Scaling Fast
Developing and deploying frontier AI models requires enormous human capital. Training large language models requires data labelers, red-teamers, safety evaluators, and infrastructure engineers. OpenAI revenue has reportedly surpassed $4 billion annually, driven by ChatGPT subscriptions, API access, and enterprise agreements.
The Rise of AI Identity Sellers
Thousands of people across the US, Philippines, India, and Kenya are now selling access to their personal identities — voices, likenesses, and conversational data — to AI companies. Platforms like Remotasks and Amazon Mechanical Turk pay individuals $5 to $50 per task to record speech, transcribe conversations, or allow facial expressions to be captured on video.
How Identity Training Works
Companies building speech recognition models need thousands of hours of recorded conversations across accents, ages, and speaking styles. Rather than relying on existing datasets, companies hire contractors to provide fresh, diverse voice data. Some participants read scripted passages; others have unstructured conversations that are transcribed and annotated.
What This Means for the AI Industry
For all its sophistication, AI still depends heavily on enormous amounts of cheap, human-generated data. The most powerful models are built not just on compute and algorithms, but on the unpaid or low-paid labor of millions of people around the world. Regulators are beginning to ask whether this new form of labor deserves the same protections as traditional employment.
FAQ
Q: How many people does OpenAI currently employ and what is the target by end of 2026?
A: OpenAI currently employs approximately 4,500 people and plans to grow to 8,000 employees by the end of 2026, adding roughly 3,500 positions across research, engineering, safety, policy, and operations.
Q: What are AI identity sellers and how do they contribute to AI training?
A: AI identity sellers are gig workers who sell access to their voices, facial data, or conversational recordings to AI companies. This data trains speech recognition, natural language processing, and video generation models.
Q: What are the privacy concerns associated with selling identity data to AI companies?
A: Privacy advocates warn that consent frameworks are often unclear, workers may not understand how their data will be used long-term, and companies holding voice and facial data could authorize third parties to create synthetic media without additional consent.