The United Nations General Assembly passed its first resolution focused on artificial intelligence. The 193 member states backing the measure aim to address the risk of AI gaining control over human decisions. A prominent AI expert and data analyst based in the United States spoke with socialbites.ca about the potential dangers of unchecked AI development and whether the resolution can curtail those risks.
“Neural networks present dangers that are hard to regulate today, illustrated by widely used systems like ChatGPT. The main concerns include the spread of information that may be harmful, particularly in dangerous areas such as illegal drugs or explosive materials. Historically, sources of information could be blocked or individuals prosecuted, but with AI, the landscape has grown far more tangled,” the expert noted.
According to the expert, ChatGPT builds a text knowledge base from vast portions of the Internet, though the exact data size remains undisclosed. The question then becomes who controls this information and how. Reliable sources covering this process are not always available.
“The second stage is the training process, which requires immense resources and time. During training, information is transformed into numerical representations, creating a neural network that encodes the patterns found across texts. It is important to understand that this is not a simple database; the original sources and decisions behind the model are not directly recoverable. This, in fact, is a central concern raised by business leaders in the field,” the expert explained.
The subsequent step is applying the trained model. Given input data, the neural network generates responses, but it becomes impossible to fully predict or control these outputs. Users can influence the questions posed and, in turn, the answers produced. Some users may attempt to bypass safeguards to access dangerous information for extreme purposes.
“Gaps in information control become evident throughout the AI chain. Recent regional developments have heightened these concerns. For instance, a major tech company has publicly announced a model that allows global continued AI development with limited oversight,” the expert stated.
According to the analyst, the resolution helps lawmakers pursue a clearer path and lay the groundwork for more understandable rules governing developers. The central idea is that UN member states collectively acknowledge the power and significance of artificial intelligence.
“This marks the first high-level document of its kind. It signals the start of broader, more detailed frameworks at local levels. The resolution emphasizes sovereignty, human rights protection, safe and risk-free use, and data privacy. It calls for a set of rules and laws governing AI research and development. It also highlights a key objective: to reduce the growing technological gap over the next decade by supporting developing countries in their AI progress. The measure is a crucial initial step in countering runaway AI and is expected to prompt many countries to adopt new legal standards and encourage tech firms to pursue enhanced technical oversight,” the expert concluded.
Earlier reports noted AI developments in South Korea and the region, which continue to shape the broader global conversation about responsible AI governance.