World model companies are keeping a lot of secrets

Billy Odell Tucker-Robinson Founder & Host — Banking With Billy Network • Intelligence Network • Data Science • AI Research • World News
Published: 2026-09-18T20:30:11.721Z • Permanent link

Google DeepMind's AlphaFold 3 has revolutionized the field of protein folding prediction, a crucial area of research in the development of new medicines. However, the model's impressive accuracy has also raised concerns about the company's data practices and potential intellectual property disputes. In a recent interview, Andrew Ng, co-founder of Coursera and former head of AI at Baidu, expressed concerns about the lack of transparency surrounding DeepMind's data collection and usage practices. Ng pointed out that the company's reliance on vast amounts of publicly available data, often obtained through unsolicited means, has raised questions about the ownership and control of that data.

Google DeepMind has been criticized for its aggressive acquisition strategies, with some companies and researchers accusing the company of using its vast resources to acquire and silence potential competitors. For example, in 2020, the company acquired a UK-based startup that had developed a protein folding model, sparking accusations that the acquisition was an attempt to stifle innovation. However, Google DeepMind has maintained that its acquisitions are aimed at enhancing its research capabilities, rather than stifling competition.

In a recent statement, Demis Hassabis, CEO of DeepMind, acknowledged that the company's growth has been rapid and that there have been instances where its data collection practices have been criticized. However, Hassabis emphasized that DeepMind is committed to transparency and has established a number of guidelines to ensure that its data collection practices are fair and respectful of users' rights.

Google DeepMind's protein folding models have the potential to revolutionize the development of new medicines, but the company's data practices have raised concerns among researchers and policymakers. For example, the company's AlphaFold 3 model has been used to predict the structure of proteins that are crucial for treating diseases such as cancer and Alzheimer's. However, some researchers have expressed concerns that the model's accuracy is not yet sufficient to justify its widespread adoption, and that the company's data collection practices are not transparent enough to ensure that the model is used responsibly.

The pharmaceutical industry is already taking notice of Google DeepMind's protein folding models, with some companies investing heavily in the technology. For example, the biotech company, Moderna Therapeutics, has partnered with DeepMind to develop new treatments for diseases such as cancer and infectious diseases. However, the partnership has also raised concerns among some researchers, who worry that the company's data collection practices may compromise the integrity of the research.

Google DeepMind's protein folding models are just one part of a larger trend towards the use of artificial intelligence in scientific research. The company's AlphaFold 3 model is part of a broader effort to develop new AI-powered tools for scientific discovery, which has the potential to revolutionize the way researchers work. However, this trend has also raised concerns about the ownership and control of data, as well as the potential for bias and errors in AI-powered models.