AIAT Blog

Generative AI has huge potential to transform society. The challenge is to harness this and other related technologies for the betterment of society and to minimise the risk of abuses.

The difficulty is that the technologies continue to race ahead of government’s capacity to establish workable regulation. I am not alone in thinking that this is one of the most important issues facing the global community.

On the 9th of December, European Union negotiators announced that agreement has been reached to establish the world’s first systemic comprehensive AI rules paving the way for laws to oversee the operation of generative AI like Chat GPT.

Systems like Open AI’s Chat GPT have exploded in use to produce human-like text photos and songs but have also raised legitimate fears about the risks to jobs, privacy and copyright protection. The recent Hollywood writers strike is one example of the challenges posed by AI.

Importantly the European Union appears to have found a mechanism to expand the regulatory coverage of AI to include ‘foundation models’ also known as ‘large language models’ which are systems trained on vast troves of written work and images sourced from the internet.

One of the biggest albeit unexpected challenges in the EU deliberations was a debate on banning public use of facial scanning due to privacy concerns with a compromise being worked out reportedly on public good principles.

The legislation will come into effect in 2025 but will generative AI have already moved on?