Will humans go extinct because of AI?  

This is the question that we commonly hear. Buried in this scenario of a takeover, superhuman AI are five assumptions which, when examined closely, are not based on any evidence. Nevertheless, these claims might be true in the future.  

The assumptions behind a superhuman intelligence arising soon are: 

• AI is getting more intelligent than us at an exponential rate 
• We’ll make its way into a general-purpose intelligence, like our own 
• We can make human intelligence in silicon 
• Intelligence can be expanded without limit 
• Once we have exploding superintelligence, it can solve most of our problems. 

Recent advances in AI have resulted in automated systems that approach and surpass human performance in fields as diverse as medicine, transportation, language and natural sciences, among others. When AI systems outperform humans in these settings, a natural question to ask is how humans will change their decision-making. Humans will likely adjust to such advancements in AI by delegating to, receiving aid from, or learning from, AI systems to improve their performance. 

Superhuman AI and human decision making 

Questions about the impact of superhuman AI on human behavior are related to the literature on cumulative cultural evolution. For example, a paper published on the PNAS page shows that there is no guarantee that human decision-making will improve response to innovations, despite the human ability to accumulate knowledge within and across generations. 

But what if innovations produced by superhuman AI result in changes in human decision-making? 

Research suggests that novelty could be a relevant factor. It has been proposed that AI systems can generate new ideas by combining familiar ideas in new ways and exploring conceptual spaces that may have been overlooked. 

The analysis 

The report analyzed professional Go players’ decisions revealing that human decision-making significantly improved following the advent of superhuman AI. They found that this improvement may be partly explained by the increased novelty in human choices made after the exogenous shock of the event.  

The study analyzed decision quality by controlling player fixed effects, and the respective periods were used as the baselines. The panels present the time trends of human decision quality from 1950 to 2021. Both panels reveal that human experts began to make significantly better decisions after the advent of superhuman AI. 

Next, the novelty in decisions was studied over the same period. The result shows a sharp increase in the yearly and monthly fixed effects on the Novelty Index following the advent of superhuman AI, suggesting that players broke away from previously observed move sequences earlier in their games. 

Conclusion 

As AI systems continue to approach or surpass human abilities, it is essential to comprehend their effects on human decision-making. The research topic, specifically within the game of Go, primarily has centered on assessing the change in human decision quality. The research is an example which shows that human decision-making significantly improved following the advent of superhuman AI, and that this improvement was associated with greater novelty in human decisions.