Amid the war of Gen AI, the tech industry is also grappling with the potential threat of AI taking over humans. With new upgrades, tech giants like Microsoft, OpenAI, Google, and Meta are working tirelessly to improve their large language models and make AI answer like humans and possibly even better. While one side of the debate fears AI taking over human jobs, Meta CEO Zuckerberg remains optimistic.
Mark Zuckerberg believes there’s no need to fear AI. In fact, he argues that while AI technology is advancing rapidly, it’s still far from posing any existential risks. During his recent interview with The Verge, Zuckerberg discussed the capabilities of his newly launched Llama3, highlighting its “state-of-the-art” performance. However, despite its improved capabilities, Zuckerberg assures us that AI is still developing and remains far from causing any threat to humans. He does, however, acknowledge the need for caution in the area of multimodality, where AI can operate across various media forms.
Interestingly, Zuckerberg’s views on AI have remained consistent throughout the years. Even before Meta announced its large language models, back in 2016, he expressed similar sentiments in an interview with Axel Springer CEO Mathias D¶pfner for a German newspaper. He stated that there’s no need to fear AI and machines, as they cannot pose a threat unless we humans train them to do so.
When asked how humans can ensure robots and computers serve us, Zuckerberg assures that by default, machines are trained to be helpful. “The default is that all the machines we build serve humans,” he says, “so unless we make a serious mistake, I believe it should stay that way.”
Zuckerberg further acknowledges that machines can outperform humans in specific tasks. However, he emphasises that this doesn’t translate to superhuman capabilities. He explains, “People build machines to do things better than humans can. There are many examples throughout history. However, just because a machine excels at one task doesn’t mean it can learn new domains or connect information and context to achieve superhuman feats. This is a crucial point to understand.”
While Zuckerberg acknowledges the potential dangers of AI, he doesn’t shy away from its potential benefits. He emphasises that the key to safe AI lies in its training. Fearing AI, he argues, will only hinder progress. “Just like we ensure safety features in other machines,” Zuckerberg says, “we need to figure out how to make AI safe as well.” He believes fearful thinking is an obstacle to real progress. After all, “self-driving cars are going to prevent car accidents. AI will be responsible for reducing one of the leading causes of death in the world. Similarly, AI systems will enable doctors to diagnose diseases and treat people better, so blocking that progress is probably one of the worst things you can do for making the world better.” Blocking this progress, Zuckerberg argues, would be detrimental to improving the world.
