Scientist and entrepreneur, Prof Gary Marcus, recently shared his views on the possible roadblocks in generative AI’s (Artificial Intelligence) future in a blog post, raising questions about its potential usefulness. He points to numerous serious and unresolved issues at the core of generative AI, including their tendency to hallucinate or create false information, and their lack of reliable interaction with external tools. This perspective has sparked discussion within the AI community, including the team at InspiredMinds, leading to deeper exploration during their Intelligent Health event slated for 13-14 September in Basel, Switzerland.
Questions brought forth involve the potential revolutionary impact of Generative AI and Large Language Models (LLMs) on healthcare, potentially driving superior outcomes. This ties into the ongoing speculation about an imminent AI revolution that could fundamentally alter the workplace. A prediction that was recently put forth by the Organisation for Economic Co-operation and Development (OECD).
However, there is also consideration of Prof Marcus’s skepticism where he highlights not just one, but a series of unsolved problems central to generative AI. The matter of generative AI producing false information, a phenomenon identified as ‘confabulation,’ stands out as a significant concern. Furthermore, these AI systems reportedly struggle to reliably interface with external tools. This combination of issues could potentially limit generative AI’s overall effectiveness and practicality.
Given the potential transformative influence of AI and LLMs, particularly in sectors like healthcare, the necessity of addressing these concerns is clear. The looming questions involve how effective these technologies will be and what, if any, are the inherent limitations? Can these fundamental issues be resolved, or are they inherently limiting factors to what could otherwise revolutionize industries?
As these discussions and debates unfold, particularly at upcoming events like Intelligent Health, the AI community is set to explore these questions further. The balance between the incredible potential of generative AI and LLMs and the critical issues pointed out by leaders like Prof Gary Marcus adds even more anticipation towards the future of this technology.