news studionews studio

尽管2023年生成式AI技术取得了巨大进步,但根据新智元最新报道,AI领域的专家们普遍认为,即使GPT-5这样的先进语言模型被推出,2024年也不会实现真正的人工通用智能(AGI)。据专家们预测,目前即便是最先进的AI模型,其智能水平也仍然有限,远远没有达到实现AGI的水平。

在2023年,生成式AI技术的发展为各行各业带来了前所未有的变革,从文本生成、图像生成到音频生成,AI的应用领域不断拓宽。然而,这些进步似乎并没有改变专家们对于AGI实现时间的看法。他们普遍认为,尽管AI技术在某些特定任务上表现出色,但在理解和处理复杂问题方面仍然力不从心。

GPT-5模型的发布可能会再次提升AI技术的水准,但据专家们分析,这只是一种“有限的突破”。他们认为,GPT-5虽然在文本生成和理解方面有更出色的表现,但其智能仍然是基于大量数据训练得来的,而非真正的“理解”。

这一观点得到了不少AI研究者的支持。他们认为,要实现真正的AGI,不仅需要AI模型有更强的学习能力,还需要具备类似人类的推理、创造和情感等能力。而这些能力的实现,可能还需要更长的时间。

英文标题:AI Experts Forecast: GPT-5 Limited, No AGI by 2024

英文关键词:AI forecast, GPT-5, AGI

News content:
Despite the significant advancements in generative AI technologies in 2023, a recent report by Xinwise suggests that AI experts widely agree that even the introduction of GPT-5, an advanced language model, will not mark the achievement of Artificial General Intelligence (AGI) in 2024. Experts predict that the current level of intelligence demonstrated by even the most sophisticated AI models is still limited, far from reaching the level required for AGI.

In 2023, the development of generative AI has brought about unprecedented transformations across various industries, ranging from text generation to image and audio generation. However, these advancements do not seem to have altered the experts’ views on the timeline for AGI. They generally believe that while AI performs excellently in specific tasks, it lacks the capability to understand and handle complex issues comprehensively.

The release of GPT-5 may again raise the bar for AI technology, but experts analyze that this represents only a “limited breakthrough.” They argue that although GPT-5 shows improved performance in text generation and understanding, its intelligence is still based on extensive data training rather than true “understanding.”

This perspective is supported by many AI researchers. They believe that achieving true AGI requires not only stronger learning capabilities for AI models but also the ability to reason, create, and experience emotions, much like humans. The realization of these abilities may still require more time.

【来源】https://www.ithome.com/0/742/812.htm

Views: 1

发表回复

您的邮箱地址不会被公开。 必填项已用 * 标注