TLDR: The ‘OpenAI to Z Challenge’ successfully leveraged artificial intelligence to identify over 100 potential archaeological sites in the Amazon rainforest. Team Black Bean emerged as the winner, utilizing deep learning with LiDAR and satellite data to reveal hidden historical information, marking a significant advancement in archaeological research.
In a groundbreaking initiative that merged cutting-edge artificial intelligence with historical exploration, OpenAI’s ‘OpenAI to Z Challenge’ has led to the discovery of numerous potential ancient sites within the vast Amazon biome. Launched in May 2025, the challenge invited participants globally to apply OpenAI’s advanced models, including o3, o4-mini, and GPT-4.1, to a wealth of open-source data to unearth previously unknown archaeological evidence.
The motivation behind the challenge was rooted in long-standing legends of the Amazon, such as the mythical ‘lost city of Z,’ Paititi, and El Dorado, which hint at dense ancient civilizations hidden beneath the rainforest canopy. Participants were tasked with sifting through high-resolution satellite imagery, published LiDAR tiles, colonial-era diaries, indigenous oral histories, and archaeological survey papers. The goal was to pinpoint new settlements, generate historical insights, or develop novel discovery methods, primarily focusing on Brazil with extensions into neighboring South American countries.
Over 10,000 individuals and teams joined the ‘OpenAI to Z Challenge,’ demonstrating widespread enthusiasm for this innovative approach to archaeology. The competition culminated in August 2025 with Team Black Bean being declared the winner. Their successful methodology involved applying deep learning techniques to public LiDAR and satellite data, creating detailed maps that effectively surfaced what lay hidden beneath the dense rainforest canopies.
Team Black Bean’s efforts reportedly identified over 100 potential archaeological sites. One team member expressed their ‘biggest wow moment’ as seeing the model’s results, which, after post-processing and manual analysis, revealed a list of potential discovery sites that aligned with archaeological knowledge and common sense. Another finalist, who developed a scalable system called AKOS, highlighted the integration of OpenAI’s intelligence to train classifiers on LiDAR and satellite images, segmenting the Amazon forest into 3×3 km tiles for detailed analysis.
This challenge has been lauded for democratizing archaeological research, making it accessible to anyone with curiosity and a structured approach, rather than being limited to specialists. OpenAI emphasized that AI empowers humanity to understand the history of millions who lived, struggled, and thrived long ago.
Also Read:
- Generative AI Workshops Drive Transformative Change in Business Operations
- UCLA Researchers Pioneer Light-Based AI for Ultra-Low Energy Image Generation
The challenge offered substantial prizes, with the first-place winner receiving $250,000 in a mix of cash and API credits, along with funding to continue their work in collaboration with archaeologists and remote sensing experts. The success of the ‘OpenAI to Z Challenge’ underscores the transformative potential of artificial intelligence in scientific discovery and cultural preservation, opening new frontiers for understanding human history.
