The Chilean government has partnered with 30 institutions across Latin America and the Caribbean to create Latam-GPT as an alternative to Big Tech LLMs, which have limited capabilities in languages other than English. Latam-GPT has collected over 8 terabytes of data from the region and is being trained by locals who take language and cultural nuances into account. The model follows Southeast Asia’s Sea-Lion, Africa’s UliaLlama, India’s BharatGPT, and Mongolia’s Egune AI. (Legal analysts have, however, issued concerns about the lack of consistent data privacy legislation in Latin America.)

Limited partners are increasingly adding AI ethics to their due diligence practices, even as ESG has fallen out of the limelight. Regulatory developments at the state level in California and New York are helping to drive the push. Early leaders include Manulife, StepStone Group, and Norway’s sovereign wealth fund.

Pharmacists in the Amazon region are using AI developed by Brazilian nonprofit NoHarm (and backed by Google and Amazon) to process prescriptions faster in overburdened public clinics. Early evidence from the project suggests it is a scalable model for under-resourced health systems.

AI “Godfather” Yoshua Bengio is launching a new non-profit AI safety research organization called LawZero that intends to prioritize safety over commercial imperatives. The organization is a response to what Bengio sees as today’s frontier models “growing dangerous capabilities and behaviours, including deception, cheating, lying, hacking, self-preservation, and more generally, goal misalignment.”

Startup Crusoe Energy Systems, currently valued near $10 billion, converts stranded fuel into computing energy to power data centers, a solution that addresses both energy inefficiency and environmental waste. Since its founding in 2018, Crusoe has prevented over 22 billion cubic feet of natural gas from being flared, which is equivalent to taking 630,000 cars off the road annually.