Last week Mark Zuckerberg announced in an open letter that Meta – parent of Facebook, WhatsApp, and Instagram – is making its AI platform Llama into an open-source tool. That means anyone can have access to the code, and build on it to make new applications.

“Open source will ensure that more people around the world have access to the benefits and opportunities of AI, that power isn’t concentrated in the hands of a small number of companies, and that the technology can be deployed more evenly and safely across society,” the ‘social network’ bazillionaire wrote.

Meta isn’t the first tech giant to open up their code, allowing everyone to build on its foundation. Golang, which we use widely at Quotech, was invented at Google as an open-source programming language. H3, which we’ve also used, is a geospatial indexing system that was invented at Uber, and is similarly open to all, and free to use.

In the world of insurance IT, the Open Exposure Data standard from Oasis is a fine example of open-source standards which allow everyone to communicate in the same digital language for free. Any other approach imposes a level of exclusivity: if you don’t pay, you cannot say. It immediately builds barriers to inter-system communication. It’s like hoping everyone will speak English, but giving dictionaries only to those who fit a certain profile (typically characterised by a willingness to pay large sums for them).

It’s disappointing, therefore, that after decades of enduring the need to nail down digital data standards for the London insurance market, an open data standard was not selected. Instead, data made fit for the Core Data Record must be parsed into a standard which is prohibitively expensive for start-up companies. The decision limits access, and therefore stymies innovation. It is a missed opportunity to get everyone involved.