UK seeks light-touch AI legislation as industry leaders call for LLM pause
- Reference: 1680111310
- News link: https://www.theregister.co.uk/2023/03/29/uk_seeks_lighttouch_ai_legislation/
- Source link:
Department for Science, Innovation and Technology has launched a [1]white paper consultation in preparation for developing legislation addressing risks inherent in deploying AI in society.
In the foreword to "A pro-innovation approach to AI regulation", science, innovation and technology minister Michelle Donelan said the risks associated with AI could "include anything from physical harm, an undermining of national security, as well as risks to mental health."
[2]
However, the British government has rejected the [3]risk-based approach taken by the EU , instead advocating a framework designed to "ensure that regulatory measures are proportionate to context and outcomes."
[4]
[5]
Donelan said: "A heavy-handed and rigid approach can stifle innovation and slow AI adoption. That is why we set out a proportionate and pro-innovation regulatory framework. Rather than target specific technologies, it focuses on the context in which AI is deployed. This enables us to take a balanced approach to weighing up the benefits versus the potential risks."
The UK government's policy position is stacked against increasing alarm that developments in large language models such as GPT-4 are running ahead of any understanding of the risks.
[6]
Hundreds of computer scientists, industry leaders and AI experts have signed [7]an open letter calling for a pause for at least six months to the training of AI systems more powerful than GPT-4.
Signatories include Apple co-founder Steve Wozniak, SpaceX, Tesla and Twitter CEO Elon Musk, New York University AI researcher and professor emeritus Gary Marcus, and Grady Booch, IEEE computing pioneer and IBM Fellow.
"AI labs and independent experts should use this pause to jointly develop and implement a set of shared safety protocols for advanced AI design and development that are rigorously audited and overseen by independent outside experts," the letter said.
[8]Europol warns ChatGPT already helping folks commit crimes
[9]Google says it did not train its AI chatbot Bard on your private emails
[10]ChatGPT, how did you get here? It was a long journey through open source AI
[11]No reliable way to detect AI-generated text, boffins sigh
“If such a pause cannot be enacted quickly, governments should step in and institute a moratorium,” the signatories said.
Others, however, were quick to point out that the LLM horse may have already bolted.
[12]
Nonetheless, a sober reminder of what might be at stake in regulating AI came with [13]news from Belgium that a man had committed suicide, apparently following a number of weeks' conversation with an online chatbot based on the open-source GPT-J language model. Belgium's digital secretary [14]said yesterday , after speaking with the family, that the situation was "a serious precedent that must be taken very seriously."
In the US, the [15]National AI Initiative Act of 2020 became law on January 1, 2021. It is designed to provide "a coordinated program across the entire Federal government to accelerate AI research and application for the Nation’s economic prosperity and national security." ®
Get our [16]Tech Resources
[1] https://www.gov.uk/government/publications/ai-regulation-a-pro-innovation-approach
[2] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=2&c=2ZCS1AR59L2dEdOZ-iWNkyAAAABg&t=ct%3Dns%26unitnum%3D2%26raptor%3Dcondor%26pos%3Dtop%26test%3D0
[3] https://eur-lex.europa.eu/legal-content/en/ALL/?uri=CELEX:52021PC0206
[4] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44ZCS1AR59L2dEdOZ-iWNkyAAAABg&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[5] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33ZCS1AR59L2dEdOZ-iWNkyAAAABg&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[6] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44ZCS1AR59L2dEdOZ-iWNkyAAAABg&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[7] https://futureoflife.org/open-letter/pause-giant-ai-experiments/
[8] https://www.theregister.com/2023/03/28/chatgpt_europol_crime_report/
[9] https://www.theregister.com/2023/03/27/ai_in_brief/
[10] https://www.theregister.com/2023/03/24/column/
[11] https://www.theregister.com/2023/03/21/detecting_ai_generated_text/
[12] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33ZCS1AR59L2dEdOZ-iWNkyAAAABg&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[13] https://www.brusselstimes.com/430098/belgian-man-commits-suicide-following-exchanges-with-chatgpt
[14] https://michel.belgium.be/fr/suicide-suite-%C3%A0-des-%C3%A9changes-avec-un-chatbot-mathieu-michel-%C2%AB-il-est-indispensable-d%E2%80%99identifier
[15] https://www.ai.gov/
[16] https://whitepapers.theregister.com/
Re: How quaint it is .....
As though HM government even has a clue what AI is. They know about creative expense claims and noses in the trough. But IT?
His Muskiness has long been of the view that runaway AI could be a huge risk to humanity, so no surprise that he's a signatory.
It would be interesting to know exactly what this bunch of boffins (what's the collective noun for boffins?) see as being so dangerous. Not that I'm suggesting that there is no danger, but the nature of that danger as they see it would be good to know. It might also be interesting (and might even have been disclosed in the open letter) how many of them are associated with AI shops that might be, say, six months behind the big guys.
Even if such a pause were to happen, voluntarily or otherwise, I'm not sure that some countries wouldn't just see it as an opportunity to catch up. This genie is out of the bottle, and isn't going back whether we like it or not.
Hmmm
A number of people, many of whom are invested in the industry, see unspecified dangers and want everyone else to stop for a bit? I'm convinced.
LLMs are going to be both more disruptive and a lot less disruptive than people are predicting. Rather like Y2K it turns out computers - on the whole - don't have as much effect on the physical world as people in the industry would like to believe. On the other hand, yes, we're going to see some real uses for the technology that will - on the whole - disrupt companies already in this space.
This current infatuation with AI keeps reminding of the evocative back story to Frank Herbert's [1]Destination: Void (the first of his Pandora series).
"In the future, mankind has tried to develop artificial intelligence, succeeding only once, and then disastrously. A transmission from the project site on an island in the Puget Sound, "rogue consciousness!", was followed by slaughter and destruction, culminating in the island vanishing from the face of the earth."
[1] https://en.wikipedia.org/wiki/Destination:_Void
Regulate ha ha ha
These donkeys couldn’t regulate themselves
What chance do they have with a program that can run anywhere in the cloud
Silly politicians
Try regulating the news paper FAILED
Try regulating social media FAILED
Have they cured crime NO
LLM FFS
Please would El Reg expand acronyms. This is just yet another example
It's a Torrie Govt
At any sign of potential cash, thoughts and concerns for people, just fcuk off.
How quaint it is .....
.... that traditional and conventional, well used levers of command and control are presumed by existing leaderships to be AI applicable and effective in confining and constraining and containing it from demonstrating novel and uncomfortably disruptive notions, undeniable true, across internetworking networks connecting operating systems/critical infrastructure.
If you think hell hath no fury like a woman scorned, you definitely haven’t crossed swords and fought to do battle against a constantly learning AI.