Last week, G7 leaders agreed upon a set of international guiding principles on artificial intelligence (AI) and proposed a voluntary code of conduct for AI developers. This development aligns with the European Commission’s ongoing efforts to regulate AI.
New advances in computing capabilities, the increasing profusion of data, and evolving algorithms have driven substantial progress in AI over the past several decades. AI involves enabling technical systems with human-like abilities including reasoning, learning, planning, and demonstrating creativity. These systems perceive their surroundings, respond accordingly, and act to attain specific objectives.
The proliferation of AI has been exemplified by technologies such as Siri, Alexa, and customer service chatbots. Last year, the development of Chat GPT underscored this trend. Chat GPT uses natural language processing to engage in human-like dialogues, allowing it to compose articles, solve mathematical problems, summarise presentations, generate titles, and answer trivia questions, among other tasks.
The series of advancements in AI over the recent years have underscored the urgent need for regulation in the field. In response, the EU Artificial Intelligence Act (“AI Act”) was drafted and is currently being finalised. This Act could make history as the world’s first comprehensive AI law.
- The G7’s initiative on artificial intelligence, referred to as the “Hiroshima process”, came into being on May 19, 2023. The initiative was launched with the objective of addressing the risks and opportunities associated with generative AI, crafting guiding principles for AI stakeholders and organisations involved in developing advanced AI, and promoting collaboration in the development of responsible AI tools and best practices.
- The Hiroshima Process’s principles emphasise identifying and mitigating risks, the importance of public reporting, robust security control investment, reliable content authentication, and the promotion of responsible information sharing.
The Hiroshima process serves to supplement the proposed EU Artificial Intelligence Act, thereby providing a comprehensive framework for the regulation of AI. The AI Act assigns responsibilities to the users and providers of AI based on the associated risks. Systems that leverage AI deemed to pose unacceptable risks will be prohibited. High-risk AI technologies will require assessment prior to their introduction into the market, as well as throughout their lifecycle. Generative technologies like Chat GPT will have to abide by transparency requirements.
Once the AI Act comes into force – anticipated by the end of 2023 – AI system providers will be given a timeframe of either 24 or 36 months to achieve full compliance.
Read more: JURIST – News