SEOUL — South Korea and Britain kick off a serious worldwide summit on synthetic intelligence in Seoul this week, the place governments plan to press tech companies on AI security.
The assembly is a follow-up to the inaugural international AI security summit at Bletchley Park in Britain final yr, the place dozens of nations voiced their fears to main AI companies concerning the dangers posed by their tech.
Security is once more on the agenda on the AI Seoul Summit beginning Tuesday and representatives are anticipated from main AI companies, together with ChatGPT maker OpenAI, Google DeepMind, French AI agency Mistral, Microsoft, and Anthropic.
“As with every new know-how, AI brings new dangers, together with deliberate misuse from those that imply to do us hurt,” South Korean President Yoon Suk Yeol and UK Prime Minister Rishi Sunak stated Monday in a joint article.
“Nonetheless, with new fashions being launched nearly each week, we’re nonetheless studying the place these dangers could emerge,” they stated within the piece, printed by the South Korean every day JoongAng Ilbo and Britain’s i newspaper.
The stratospheric success of ChatGPT quickly after its 2022 launch sparked a gold rush in generative AI, with tech companies all over the world pouring billions of {dollars} into growing their very own fashions.
READ: AI consortium to make sure AI security worldwide
Generative AI fashions can generate textual content, pictures, audio, and even video from easy prompts, and its proponents have heralded them as a breakthrough that can enhance lives and companies all over the world.
However critics, rights activists, and governments have warned that they are often misused in all kinds of conditions, together with the manipulation of voters by pretend information tales or so-called “deepfake” footage and movies of politicians.
Dramatic adjustments
Many have known as for worldwide requirements to control the event and use of AI.
“After we meet with firms on the AI Seoul Summit, we are going to ask them to do extra to point out how they assess and reply to threat inside their organizations,” Yoon and Sunak wrote.
“We may even take the subsequent steps on shaping the worldwide requirements that can keep away from a race to the underside.”
The Seoul summit comes days after OpenAI confirmed that it had disbanded a workforce dedicated to mitigating the long-term risks of superior AI.
READ: AI will probably be in every single place in 2024
The 2-day summit will probably be partly digital, with a mixture of closed-door classes and a few open to the general public in Seoul.
Nonetheless, a bunch of six South Korean civil society organizations, together with the distinguished Peoples Solidarity for Participatory Democracy, criticized the summit’s organizers for not together with extra growing nations.
“It could be useful to debate worldwide norms for AI in a extra open discussion board the place all international locations and numerous stakeholders from all over the world can take part equally, somewhat than in an elite membership of some developed international locations,” they stated in a joint assertion on Monday.
Along with security, the summit will talk about how governments might help spur innovation, together with into AI analysis at universities.
READ: Amazon, Google, Meta, Microsoft, different tech companies comply with White Home-set AI safeguards
Contributors may even take into account methods to make sure the know-how is open to all and might help in tackling points corresponding to local weather change and poverty.
“It’s simply six months since world leaders met at Bletchley, however even on this quick house of time, the panorama of AI has modified dramatically,” Yoon and Sunak stated.
“The tempo of change will solely proceed to speed up, so our work should speed up too.”
Your subscription couldn’t be saved. Please strive once more.
Your subscription has been profitable.
France will host the subsequent AI security summit.
