The Chief AI Scientist at Meta, Yann LeCun, recently addressed the capabilities and limitations of artificial intelligence (AI) systems, shedding light on the ongoing debate surrounding the technology’s risks. During a panel discussion at the Viva Tech conference, LeCun emphasized that current AI systems, like ChatGPT developed by OpenAI, lack human-level intelligence and are merely on par with the intelligence of dogs.
LeCun explained that AI systems like ChatGPT are large language models trained on vast amounts of text data, enabling them to respond to user prompts in a language understood by humans. However, he stressed that these systems have limited understanding of the real world since they solely rely on language and lack comprehension of underlying realities beyond text.
The Meta AI chief acknowledged that AI’s impact on society hinges on its application. While cautioning against the development of AI for harmful purposes such as weapons or fossil fuel production, LeCun highlighted the positive potential of AI in areas like healthcare, education, and culture.
Addressing the future trajectory of AI, LeCun underscored the need to train AI systems using more than just language. He mentioned Meta’s efforts to train AI on video, acknowledging the challenges associated with this approach.
LeCun also mentioned the disparities between AI and human cognitive abilities. While AI systems may pass professional exams like the Bar, they still struggle with simple tasks such as loading a dishwasher, which a 10-year-old child can easily learn. He concluded that achieving even the intelligence level of dogs would require significant advancements in AI.
Regarding concerns about robots overtaking humanity, LeCun dismissed the notion, asserting that machines with superior intelligence should be seen as beneficial and supportive assistants, rather than threats. He stressed the importance of ensuring AI systems remain controllable and subservient to humans.
The discussion also touched upon the ethical and regulatory aspects of AI. Jacques Attali, a French economic and social theorist, raised the question of who should be responsible for establishing boundaries in AI development, emphasizing the need for safeguards.
As the debate surrounding AI’s potential risks and benefits continues, experts like LeCun and Attali highlight the importance of ethical considerations, regulation, and the responsible application of AI technologies in shaping a better future.
