The responsible development and deployment of AI has become a topic of intense debate in recent times. Headlines about AI suggesting people should eat rocks or AI-generated beauty contests have sparked discussions about the risks and flaws associated with AI. However, while there are valid concerns about the potential harm from AI tools like deepfakes and job displacement, there is no evidence to suggest that AI is inherently out to harm or kill us.
Despite the lack of concrete evidence, 13 current and former employees of leading AI providers recently issued a whistleblowing letter, expressing grave concerns about the risks AI poses to humanity. The fact that these whistleblowers include experts who have worked closely with cutting-edge AI systems adds weight to their concerns. However, it is worth noting that the letter did not contain any jaw-dropping allegations, which could be due to restrictions imposed by their employers or the scarcity of evidence supporting their worries.
One undeniable fact is that generative AI models continue to improve in terms of their capabilities. With advancements in scaling models and training them on larger datasets, it is widely accepted that these models will become even smarter in the next year or two. In fact, renowned AI researcher Geoffrey Hinton believes that artificial general intelligence (AGI), which matches or exceeds human-level intelligence across most cognitive tasks and domains, could be achieved within five years. This is a significant shift in Hinton’s viewpoint, as he previously believed AGI was decades away.
However, not everyone agrees with the optimistic projections for AGI. Some experts, like AI influencer Gary Marcus, question the scalability of these models and speculate that we may be witnessing early signs of a new “AI Winter” instead. This refers to a period of disillusionment and criticism when AI fails to deliver on overly ambitious promises, leading to declines in interest and funding.
Recent trends in generative AI dealmaking indicate a decline in investment deals and size. If this trend continues, it could lead to existing companies becoming cash starved and potentially ceasing operation. Additionally, a Fast Company article suggests that there is little evidence of AI technology unleashing enough new productivity to boost company earnings or stock prices, which could further contribute to the rise of a new AI Winter.
Despite the potential challenges and uncertainties surrounding AI, many believe that its transformative potential cannot be ignored. Gartner compares AI to the printing press and electricity, emphasizing its potential to impact society on a large scale. The belief in AI’s transformative power is supported by evidence of advanced gen AI models outperforming humans in certain tasks, such as persuasion and emotional support.
The question of whether AI will solve our greatest challenges or destroy humanity remains unanswered. It is likely that there will be a mix of positive advancements and regrettable harm stemming from advanced AI. Tech billionaires themselves, like Elon Musk and Mark Zuckerberg, are divided on the potential risks and benefits of AI. Responsible AI development, ethical principles, safety testing, human oversight, and robust control measures are essential to navigate this rapidly evolving landscape.
While the future of AI remains uncertain, it is crucial to engage in informed dialogue, prioritize ethical development, and establish proactive oversight to ensure that AI benefits society. The dreams of a world of abundance and leisure could become a reality, but without careful consideration and responsible development, they could also turn into a nightmarish hellscape.