Sources and documents: Google asks employees to test potential ChatGPT competitors, including Apprentice Bard, a chatbot that uses its LaMDA conversational tech
- Google is testing ChatGPT-like products that use its LaMDA technology, according to sources and internal documents acquired by CNBC.
Context & Ripple Effects
Google’s internal testing follows its reported “code red” response to ChatGPT and a period in which the company was reportedly reluctant to release LaMDA because of reputational concerns. The employee program turns that internal tension into a product-evaluation process around Apprentice Bard.
The near-term arc is toward public availability: Google subsequently introduced Bard as an experimental conversational service. That makes the testing significant as a bridge between LaMDA research and a direct ChatGPT competitor.
First-order effects
- Google employees become the first evaluation group for Apprentice Bard and other LaMDA-based conversational products, giving Google internal feedback before broader release decisions.
- Apprentice Bard moves Google’s response to ChatGPT from executive concern and team reassignment into active product testing.
Second-order effects
- Google’s move raises pressure on other named ChatGPT alternatives—including DeepMind’s Sparrow, Anthropic’s Claude, and Character AI—to differentiate as Google prepares its own conversational offering.
- Google’s eventual Bard debut makes speed-to-release a more central competitive variable alongside the safety and reputation concerns that had delayed LaMDA’s rollout.
Third-order effects
- If internal testing becomes the standard path to launch, leading AI labs will increasingly pair frontier conversational models with controlled-release programs rather than treating model capability alone as the product.
- The episode points to conversational AI becoming a strategic interface battle among major platforms, with deployment discipline and public trust shaping which models reach users.
The trend: Conversational AI is shifting from a research and reputation question into a platform-release race, with controlled testing mediating the move to public products.