Jim Thomas, Scan the Horizon
Una de las características más sorprendentes de la biología sintética impulsada por la IA es que gran parte del trabajo está siendo dirigido por las mayores empresas tecnológicas del mundo. La mayoría de estos nuevos líderes de IA/SynBio son titanes digitales, sin experiencia previa en biotecnología o en la gestión de la biodiversidad, pero con una amplia experiencia en la implantación de modelos de negocio monopolísticos y en eludir las normativas. Están firmando acuerdos conjuntos o adquiriendo empresas biotecnológicas más pequeñas.
Esto es relevante para los debates sobre Synbio y la evaluación de riesgos. Pero también es relevante para la DSI, porque estas bases de datos, conjuntos de datos y servicios privados utilizan enormes cantidades de DSI, y las empresas que los gestionan pretenden ganar dinero con ellos.
Google/Alphabet – Google DeepMind (AI research laboratory) developed the high profile Alphafold program. Google has a joint venture with leading SynBio company Gingko Bioworks to generate novel proteins. They have also created their own biotechnology company called Isomorphic Labs, which is using AI to generate new drug compounds for major pharmaceutical companies.
Microsoft – The CEO of Microsoft AI, Mustafa Suleyman, recently published a high-profile book (called The Coming Wave) on how the convergence of SynBio with AI will transform society (and create new risks). His firm is developing several generative AI tools for SynBio, including a generative medical platform called BioGPT.
Amazon – In June 2024, the world’s largest provider of data cloud services announced it was collaborating with a company called EvolutionaryScale to host ESM3 – a generative biology AI platform trained on “billions of protein sequences spanning 3.8 billion years of evolution”. According to Amazon, ESM3 can understand complex biological data from various sources and generate entirely new proteins that have never existed in nature. Meanwhile, the Bezos Earth Fund (associated with Amazon founder Jeff Bezos and his girlfriend) has launched a 100 million dollar ‘AI for Climate and Nature’ program that focuses on using generative AI for alternative proteins and other materials. Amazon is reportedly also interested in brain organoid biocomputation.
NVIDIA – The world’s largest AI chipmaker is also out front in generative biology. Their GenSLM AI platform has been trained on hundreds of thousands of genomes to generate novel microbial and viral genomes. They have particularly been developing candidate COVID sequences for vaccines and appear to have successfully
predicted some new Covid variants. Nvidia is also collaborating with Amazon’s ESM3 platform.
Meta (Facebook) – In 2022, Meta unveiled its ESMfold platform as a rival to Google’s Alphafold. Meta claimed that ESMfold housed more than 600 million protein structures and was 60 times faster than Alphafold. However, the project was shut down in 2023 amidst large staff layoffs at Meta.
Salesforce – This leading US data cloud company has developed ProGEN, an AI large-language model for gene-rating novel proteins. ProGEN was trained by feeding the amino acid sequences of 280 million different
proteins into a machine-learning model. As a proof-of-concept, Salesforce then tuned the model by priming it with 56,000 sequences from just one class of protein: lysozymes (used for food ingredients).
Alibaba – In 2023, scientists at the leading Chinese technology giant, Alibaba, published results from its LucaProt AI platform, which was trained to identify RNA viruses. According to the researchers, LucaProt identified 161,979 potential RNA virus species and 180 RNA virus supergroups. They asserted, “This study marks the beginning of a new era of virus discovery, providing computational tools that will help expand our understanding of the global RNA virosphere and of virus evolution.”
For sources, check the report “Black Box Biotechnology”