EU, US To Seek Stopgap Standards for AI, EU Tech Chief Says (reuters.com) 8
The European Union and the United States are set to step up cooperation on artificial intelligence with a view to establishing minimum standards before legislation enters force, the EU's tech chief Margrethe Vestager said on Tuesday. From a report: The European Union's AI Act could be the world's first comprehensive legislation governing the technology, with new rules on facial recognition and biometric surveillance, but EU governments and lawmakers still need to agree a common text. Vestager, a vice-president of the European Commission, told a briefing on Tuesday that process might be completed by the end of the year.
"That would still leave one if not two years then to come into effect, which means that we need something to bridge that period of time," she said. Vestager said AI would be one area of focus at the fourth ministerial-level meeting of the Trade and Technology Council (TTC) in Sweden on May 30-31, with discussions on generative AI algorithms that produce new text, visual or sound content, such as ChatGPT. "There is a shared sense of urgency. In order to make the most of this technology, guard rails are needed," she said. "Can we discuss what we can expect companies to do as a minimum before legislation kicks in?"
"That would still leave one if not two years then to come into effect, which means that we need something to bridge that period of time," she said. Vestager said AI would be one area of focus at the fourth ministerial-level meeting of the Trade and Technology Council (TTC) in Sweden on May 30-31, with discussions on generative AI algorithms that produce new text, visual or sound content, such as ChatGPT. "There is a shared sense of urgency. In order to make the most of this technology, guard rails are needed," she said. "Can we discuss what we can expect companies to do as a minimum before legislation kicks in?"
broad qualitative statements in standards or law (Score:1)
Re:broad qualitative statements in standards or la (Score:4)
Since the real goal is regulatory capture by the current players in the game, that may be entirely by design.
Re: (Score:3)
I mean, how do they define "AI" enough? and how can they tell a private company what they can/cannot develop in house. Do the rules ONLY apply when the AI is exposed publicly?
Does this only apply to companies that are a certain size? Does this apply to the private citizen working in the proverbial garage?
Does this apply to people that expose it to the public for $$ or what about if it is open to the public for free?
I mean, this is sof
Re: (Score:2)
makes sense (Score:3)
Totally makes sense.
Put some rules in place that the hyped-up AI companies can use as scapegoats before people realize that ChatGPT, for all the cool show effect it has, is essentially just autocorrect on steroids.
There was a cool paper published recently that demonstrated how the "quantum leap" and "emergent properties" of the current models are the result of carefully chosen metrics and disappear when the data is analyzed properly. AI is simply a matter of scale. Give it more data and more crunching power and it'll perform better, if I recall correctly in a fairly linear manner. We knew that 50 years ago when neuronal nets were the hype the first time around, we just didn't have the computing power to make it happen.
So yes, regulate it. Otherwise the hype will implode and a couple billions of "value" will evaporate.
Re: (Score:2)
Totally makes sense. Put some rules in place that the hyped-up AI companies can use as scapegoats before people realize that ChatGPT, for all the cool show effect it has, is essentially just autocorrect on steroids.
Addendum: it turns out that "autocorrect on steroids" is enough to replace a significant portion of the workforce, and is also a fairly good approximation of how we interact with most folks most of the day.
THis Will Be Addressed (Score:2)