UK market authority lays down guidance for principled use of AI
Helping you help us all.
For good or ill, artificial intelligence is here to stay. That's actually underselling it because the AI industry is booming. Just ask Nvidia. Industries across the world are spending billions looking at ways to integrate AI models into their businesses. But while AI has the potential to boost productivity, accelerate innovation and become a useful tool for end users, global policy has yet to catch up to meet increasing concerns.
The UK Competition and Markets Authority (CMA) has released a report following a review of Foundation Models (FMs). It outlines a series of principles that it hopes will lead to measures that protect consumers, and foster healthy competition and responsible AI development.
The seven key guiding principles outlined in the report are accountability, access, diversity, choice, flexibility, fair dealing and transparency. They're all laudable goals. But it's that last one is one that needs to be highlighted. Can morals and ethics be broken down into ones and zeros?
Sarah Cardell, CEO of the CMA, said: “The CMA’s role is to help shape these markets in ways that foster strong competition and effective consumer protection, delivering the best outcomes for people and businesses across the UK".
These are important goals, and the CMA deserves credit for its proactive approach. But the report mentions it is limited in scope. Key questions including the protection of intellectual property, the spread of misinformation, the potential for fraud, the need for data protection and security have not yet been addressed. And that highlights just how far there is to go to ensure AI models don't become tools for nefarious purposes.
The AIs the public are most familiar with are Large Language Models (LLMs) like ChatGPT. They're essentially a huge database of data scraped from the internet, and not all of that data is legitimate. Some of it may be proprietary, some of it may be illegal, and some of it is just plain garbage. A comprehensive regulatory framework is needed, but that's sure to be easier said than done.
Just about every day we see news stories saying someone lost their life savings. Scams and fraud is rife. Malware, ransomware and things like identity theft are all prevalent. Some rogue AI scambot could be used for all kinds of shitty purposes. Hopefully the big players in the industry are doing more than just talking about ensuring their models cannot be used for villainy. You just know that scumbags responsible for the scourge of malware are looking at ways to leverage AI, damn them to hell.
It's a new world out there, and lawmakers need to get out ahead of it. Easier said than done I suppose, but it is good to see the CMA taking a proactive approach. Now all we need is some protection from global extinction!
The AI that impersonates Jesus is one I'll steer well clear of.