ORIGINAL LINK : https://www.etoday.co.kr/news/view/2299712
Though The Blechley Declaration from the UK is Significant
The Essence of International Response is Competition, Not Cooperation
Must Strengthen Capabilities Ahead of Next Year’s Mini Conference
Yesterday, a report from Japan emerged stating that a fake video of Prime Minister Fumio Kishida is causing social confusion. The video, manipulated using Artificial Intelligence (AI) technology, appears to show Prime Minister Kishida making inappropriate comments. Uploaded to the former Twitter, known as X, on the 2nd, it garnered over 2.3 million views in just one day. At the same time, the British band The Beatles’ latest song “Now and Then” was released on YouTube. The new song is a restoration of a poor-quality demo recorded by John Lennon (1940~1980) in 1977, cleaned up using AI technology. On its first day of release, it easily surpassed 4 million views.
If we compare AI to a shaggy dog, then the Kishida video would be the sharp fangs at the front, and The Beatles’ new song “Now and Then” would be the comforting tail at the back. It’s easy to distinguish between the two, and appropriate responses are possible, as demonstrated by Nihon TV’s stern reaction to the misuse of its news program logo in the Kishida video. However, when the fur grows thick enough to cover the entire body, it could pose a significant problem, starting with difficulty in distinguishing between the front and the back.
On the 1st and 2nd of this month, representatives from 28 countries gathered at Bletchley Park in the UK to pledge international cooperation to prevent catastrophic damage that could be caused by AI. The first ‘AI Safety Summit’ took place, with President Yoon Seok-yeol participating via video. A joint declaration was also issued, known as ‘The Bletchley Declaration.’ UK Prime Minister Rishi Sunak described it as a landmark agreement. It was the first summit-level meeting to acknowledge that the shaggy dog of AI could potentially cause accidents, so it is fair to say that it was a significant event.
The UK government explained through a statement that the representatives of each country and big tech companies recognized the importance of cooperating to test the safety of next-generation AI models. They also plan to publish a report for understanding the risks of AI. However, it would be a mistake to think that the international community will move towards AI cooperation as a result of this meeting. It is closer to the truth to understand it as a reconnaissance battle aimed at leading AI regulation.
The world has been in turmoil since the unveiling of OpenAI’s ChatGPT in November last year. The United States, which owns several big tech companies, is busy containing China, while Europe, which doesn’t have any significant tech companies, is betting on regulation. The European Parliament, the legislative body of the European Union (EU), passed a draft regulation in June that includes the obligation to prevent the creation of illegal content. However, President Joe Biden of the United States signed an executive order on ‘Developing and Using Safe and Reliable AI’ on the 30th of last month. It seems like the U.S. has taken the fast lane.
It is noteworthy that the United States and the UK declared ahead of this meeting that they would lead the AI Safety Research Institute. China advocates regulation through international regulatory agencies. It’s every man for himself. If you didn’t hear the sound of the game pieces being shuffled at this meeting, you must be hard of hearing. The wrangling between early and late entrants further confuses the situation. This is evidence that a kind of standard competition is taking place.
In order not to fall behind in the AI trend, we must not only point out the weaknesses of international trends but also scrutinize whether AI regulation can be effective and if there are alternatives with a keen eye. The most likely regulatory targets that the international community can easily agree on are naturally the shaggy front parts, including the Kishida video type. However, making AI observe legal and moral limits could be a more difficult technical challenge than expected.
It is also doubtful whether holding AI platforms responsible will be effective. Can we really prevent malfunction by pressuring companies while ignoring user misuse? There is a chance that rogue states, terrorist organizations, and criminal groups could play with fire.
Albert Einstein, who changed the 20th-century world with the theory of relativity, is said to have quipped, “It’s enough if he knows himself,” when a friend complained that his pet dog Moses’s fur was so long that it was hard to tell front from back.
However, human society cannot be so nonchalant in the face of the unpredictability of AI. Korea will host an AI mini-summit in May next year with the UK to review follow-up measures. It’s a bold move to get ahead in the competition for AI leadership. We must have the insight and foresight to see how wide the danger radius of AI is. We need to pour all our energy into enhancing our ability to identify the front and back of the AI shaggy dog and control its dangers.
댓글0