Artificial Intelligence (AI) Safety Summit: A Mission Impossible? 28 states signed the Bletchley Declaration, a joint statement that acknowledges the risks

ΣΧΕΤΙΚΑ ΑΡΘΡΑ

Ο Πρωθυπουργός πάει στο Ηνωμένο Βασίλειο, την επόμενη εβδομάδα, για να δει τον Ρίσι Σουνάκ: Θα ζητήσει από τον Βρετανό την επιστροφή των Γλυπτών του Παρθενώνα

Οι πολλαπλές εφαρμογές της τεχνητής νοημοσύνης: Eίναι πλέον το νέο δεδομένο στην καθημερινότητά μας

Κατακλυσμό fake news και επικοινωνιακό χάος ελέω τεχνητής νοημοσύνης φοβούνται προεκλογικά στις ΗΠΑ

By ROXANNE KAPANTZAKI

The latest action-packed movie “Mission Impossible: Dead Reckoning Part I” presents the perils of an Artificial Intelligence (AI) software that develops a will of its own. Much as this is a science fiction script, it still bears resemblance to the way our world will be shaped in the future unless the states take considerable steps against it.

Bearing in mind the fears and concerns of international community regarding the AI, Rishi Sunak took the initiative to organize a summit in the United Kingdom in order to discuss the problems arising from the technological advances in AI.

The summit took place in Bletchley Park where Alan Turing and other codebreakers helped the Allied Forces decipher the Nazi’s Enigma Code and defeat them.[1]

• Symbolising as it may be, it is rather doubtful whether this place shall prove to be a point of unity to break the secrets of the advances in the AI and therefore, control it. The question lingering is whether this modern “Enigma” Summit was successful or not and what its wider impact is.

To begin with, the British Prime Minister, Rishi Sunak, aimed at bringing together actors from all over the world, not only politicians but businessmen and companies as well that are associated with the field of Artificial Intelligence (i.e. OpenAI, Google and Meta[2]).

His ulterior motive was to make the United Kingdom a hub for AI safety and prevent criminals and terrorists from using it in order to create weapons of mass destruction.[3] In fact, the UK attempts to establish a common ground on an international level that sets out some principles “on the ethical and responsible development of AI models”.[4] It aims at the cooperation of the UK and the other states so as to apply safeguards to the technology.[5]

It is true that there are negative aspects of the AI that should be harnessed. These include the fact that this technology can generate disinformation, misidentify people due to the colour of their skin, spread election misinformation and even make discriminatory decisions concerning the illegible candidates for a job.[6]

However, the various states that participated in the Summit did not only take these aspects into consideration, but had other underlying goals. For example, the United Kingdom through this Summit intends to get a leading role in AI talks on a global level and uses this kind of soft power in order to reinforce its position as a global actor in the field of Artificial Intelligence and international relations in general.

Kamala Harris, the American Vice President, highlighted the creation of the AI Safety Institute that is about to develop all the indispensable means in order to contain the perils originating from the AI and cooperate with other international institutions including the British AI Safety Institute. Reading behind the lines, one can sense the will of the Americans to guide the international agenda concerning the risks of the AI. In fact, Kamala Harris also stressed out that the “domestic AI policies will serve as a model for global policy”.[7] In other words, America shares the same ambitions as the United Kingdom.

There is also a third actor in this field, aspiring to lead the way in AI and that is China. Many critics were against its presence in the AI Summit, but Elon Musk held another point of view. In an event organized in Lancaster House, he pointed out the importance of harmonizing the acts of the USA, the UK and China in the field of Artificial Intelligence. In addition, he also claimed that it is of paramount importance to have China in this Summit as its role is “[..] essential, really. If they’re not participants, it’s pointless”.[8]

With regard to the impact of the UK’s AI Summit, there were 28 states that signed the Bletchley Declaration, a joint statement that acknowledges the risks of AI, America and Britain expressed their will to launch their own AI safety institutes and in the near future there are two more summits, one in South Korea and another in France that are going to take place.[9] China and the USA, geopolitical rivals, managed to get involved in discussions concerning a field of wild competition and this was no small feat.

• Nevertheless, the issue of who will lead the efforts of the states and how their goal is going to be achieved remains a thorny problem. The participants emphasized the need to deter frontier AI (that is the most advanced systems) from getting out of control.

There were also 11 signatories that agreed to have their AI models tested before and after their public release. China did not sign this agreement but the EU, the USA, the UK, Australia and Japan did. Regardless of this agreement, the fact that there is no regulation to render it binding casts doubt on its implementation. It is a void that the EU’s AI Act and the USA’s executive order on AI use will certainly seek to fill.[10]

Generally, the AI summit in London paved the way to discussions that will alleviate the dangers of AI on an international level, underlining the UK’s ambitions to be the leading power in launching talks on AI, but if these talks do not become legally binding and the conflicting interests of the various states are not resolved, then harnessing the AI may prove to be a hard task or a ‘Mission Impossible’ indeed!

[1] B. Perrigo, “U.K.’s AI Safety Summit Ends With Limited, but Meaningful, Progress”, 2 Nov 2023, online at: https://time.com/6330877/uk-ai-safety-summit/, (accessed 11/11/2023).

[2] M. Bergen, “UK Tries to Bring China to the Table at AI Summit”, 1 Nov 2023, online at:   https://www.bloomberg.com/news/newsletters/2023-11-01/uk-ai-safety-summit-rishi-sunak-invites-china#xj4y7vzkg, (accessed 11/11/2023).

[3] M. Coulter, “Britain aims for global leadership role with AI safety summit”, 18 Oct 2023, online at: https://www.reuters.com/technology/britain-aims-global-leadership-role-with-ai-safety-summit-2023-10-18/, (accessed 11/11/2023).

[4] R. Browne, “The UK is gearing up for a pivotal summit on AI. Here’s what you need to know”, 31 Oct 2023, online at: https://www.cnbc.com/2023/10/31/what-you-need-to-know-about-uk-ai-summit-attendees-agenda-and-more.html, (accessed 11/11/2023).

[5] Ibid

[6] J. Bhuiyan, “How the UK’s emphasis on apocalyptic AI risk helps business”, 31 Oct 2023, online at: https://www.theguardian.com/technology/2023/oct/31/uk-ai-summit-tech-regulation, (accessed 11/11/2023).

[7] M. Gannon, “In U.K., VP Harris says AI could do ‘profound good’ or ‘endanger the very existence of humanity’”, 1 Nov 2023, online at: https://ny1.com/nyc/all-boroughs/news/2023/11/01/vp-harris-delivers-remarks-on-ai-in-the-u-k–where-she-is-attending-summit, (accessed 11/11/2023).

[8] A. Smout, “Elon Musk and Rishi Sunak chat China, killer robots and the meaning of life”, 3 Nov 2023, online at: https://www.reuters.com/technology/elon-musk-says-good-uk-us-china-align-ai-safety-2023-11-02/, (accessed 11/11/2023).

[9]  M. Coulter and P. Sandle, “AI summit a start but global agreement a distant hope”, 6 Nov 2023, online at: https://www.reuters.com/technology/ai-summit-start-global-agreement-distant-hope-2023-11-03/, (accessed 11/11/2023).

[10] D. Milmo, “TechScape: Why Sunak’s ‘vanity jamboree’ on AI safety was actually…a success”, 7 Nov 2023, online at: https://www.theguardian.com/technology/2023/nov/07/techscape-rishi-sunak-ai-safety-summit-bletchley-park, (accessed 11/11/2023).

50th Polytechnic uprising anniversary march concludes