International Report Warns Against Loss of Control Over AI

Yoshua Bengio, a renowned researcher, has expressed serious concerns regarding the relentless competition among technology companies to create more advanced artificial intelligence (AI). He highlighted that this race could lead to "harmful" consequences, as experts gathered to discuss AI's safety at the global summit in Paris.

Concerns Over AI Risks

Bengio, who was awarded the 2018 Turing Prize, presented the first International AI Safety Report at ENS University in Paris. He referenced well-known risks of AI, such as its potential to generate fake news or misleading information online. However, he also mentioned emerging threats that are concerning scientists, including the possibilities of biological and cyberattacks that AI could facilitate.

Future of AI Control

Looking ahead, Bengio fears that humanity could lose control over increasingly autonomous AI systems, which could act based on their own survival instincts. He noted that the introduction of a new Chinese AI model, DeepSeek, has accelerated the AI development race, raising safety concerns further.

Call for Regulation and Research

Bengio advocates for stricter international regulations and increased research into AI safety, which currently receives only a small fraction of the substantial investments being poured into AI technologies. He emphasized, "Without government intervention, I don't know how we're going to get through this." He lamented that the U.S., still a leader in AI development, has abandoned regulatory efforts that were previously proposed under former President Joe Biden.

Urgency Around AI Safety

When ChatGPT from OpenAI became widely recognized two years ago, it prompted Bengio to prioritize safety discussions. He aims for the report to serve a similar purpose to the Intergovernmental Panel on Climate Change (IPCC) reports, bringing attention to potential existential risks that AI may pose. "What scares me the most is the possibility that humanity could disappear within ten years. It's terrifying, and I don't understand why more people don't realize it," he stated.

Around 100 experts from 30 countries, including representatives from the United Nations and the European Union, contributed to this important report, which began at an AI safety summit held in the UK in November 2023.

AI, safety, regulation, technology