Nicholas Dirks, the esteemed leader of The New York Academy of Sciences, recently took the stage at a highly acclaimed international tech gathering to share his insights on the intersection of artificial intelligence (AI), culture, and language.
His talk, entitled ‘The Social Life of AI,’ was delivered at the LEAP conference at the Riyadh Exhibition and Convention Center. Within the broader scope of DeepFest, a segment dedicated to AI that attracted over 150 speakers and an audience nearing 50,000, Dirks explored the historical evolution of human-machine interaction, beginning with the industrial revolution and extending into the profound societal changes anticipated with the advancement of AI.
Dirks articulated a vision that extends beyond the dichotomy of technological optimism and existential concerns. ‘I aim to shed light on how AI will permeate our daily existence, influencing aspects that transcend the obvious significance of technology in our economic and political spheres,’ he commented.
The heart of Dirks’ presentation was the exploration of sophisticated language models, like ChatGPT and predictive text features, which harness machine learning to generate coherent and contextually appropriate language. He underscored the critical role of cultural and contextual awareness in the programming of these AI systems.
‘Our lives are enmeshed in networks defined by cultural and historical nuances; some are universal, while others are distinctly particular to certain groups or situations. Language is pivotal, but so is the cultural backdrop within which language is curated, acquired, and utilized,’ Dirks elaborated.
Creating Social Meaning
Dirks warned of the potential cultural dislocations that unregulated AI might precipitate, potentially endowing machines with undue autonomy in the construction of social meaning, and eroding human agency. This, he suggested, could lead to misplaced trust in machines over humans. He pointed to the societal challenges already emerging from technological shifts, such as job displacement due to automation, mental health issues tied to social media algorithms, and the spread of falsehoods via deepfakes.
To mitigate these risks, Dirks advocated for a collaborative approach to AI development, calling on experts from diverse fields such as the humanities, social sciences, psychology, and philosophy to contribute. He promoted the idea of embedding AI technologies with socially beneficial uses from the outset.
Concluding on a hopeful note, Dirks emphasized the potential for a harmonious coexistence between AI and human social structures. ‘By engaging thoughtfully across sectors and disciplines, we can shape AI to enhance our lives, ensuring it complements rather than dominates or sidelines us, and that its advantages are equitably distributed,’ he concluded.