Meta and Google release computer code for chatbots.
Last year, Meta was criticised for releasing poorly understood or dangerous technology when it shared raw computer code for building a chatbot. But now, Google has also released the computer code which powers its online chatbot, indicating that sharing A.I. technology is becoming popular among industry counterparts.
Google aims to re-engage third-party developers and make Google-based A.I. models an industry standard for modern A.I. building. While Google has no plans to release its flagship A.I. model, Gemini, it has started charging for access to the most powerful version.
Many companies are afraid that A.I. technologies can be used to spread disinformation, hate speech, and other toxic content. Some are becoming secretive about the methods and software that underpin their A.I. products, while others believe that freely sharing such technology allows outsiders to identify and solve problems.
Meta’s chief A.I. scientist, Yann LeCun, believes that A.I. should be outside the control of powerful companies like Google, Microsoft and Meta, in order to be embraced by consumers and governments. OpenAI’s competitive pressure led Google to become more secretive about its A.I. technologies, but due to developer interest, it has decided to make its A.I. more freely available again.
Google is allowing people to download systems that have been trained on enormous amounts of digital text from the internet, but this typically requires significant resources that most organisations and individuals do not have.
