Karen Hao on the Empire of AI, AGI evangelists, and the cost of belief

In the world of artificial intelligence, a powerful ideology drives the relentless pursuit of artificial general intelligence (AGI), which advocates for advancements purportedly aimed at benefiting humanity. This fervor for AGI is epitomized by OpenAI, a leading entity in the field, which fervently promotes its vision of a future where AI enhances human life. Karen Hao, a journalist and author of 'Empire of AI,' recently shared insights on this phenomenon in a conversation with TechCrunch. She likens the current AI landscape, particularly the role of OpenAI, to an empire, asserting that its influence extends beyond mere economic power to significant political sway. Hao emphasizes that OpenAI's actions have redefined global dynamics, stating, 'They’re terraforming the Earth. They’re rewiring our geopolitics.' OpenAI describes AGI as a sophisticated system capable of surpassing human performance in various economically valuable tasks. This vision has ignited a surge in the AI sector, leading to heightened demands for resources and data, while raising concerns about the environmental impact and ethical implications of releasing untested technologies into society. Experts warn that the anticipated benefits of AGI may never materialize, as the industry prioritizes speed over caution. Hao argues that the current trajectory of AI development is not the only path forward. She advocates for the exploration of new algorithmic techniques that could yield advancements without the extensive resource consumption currently associated with the industry. However, the prevailing mindset within OpenAI and other tech giants emphasizes rapid progress, often at the cost of safety and ethical considerations. The financial stakes in AI development are staggering. OpenAI projects a cash burn of $115 billion by 2029, while competitors like Meta and Google are also committing tens of billions to AI infrastructure. Despite these colossal investments, the promised 'benefits to humanity' remain elusive, as societal harms such as job displacement and mental health issues continue to accumulate. Hao highlights the plight of content moderators in developing countries, who are often exposed to disturbing material while receiving minimal pay. She contends that the narrative of prioritizing AI advancements over present-day harms is misleading, advocating instead for solutions like Google DeepMind's AlphaFold, which has made significant strides in drug discovery without the associated ethical dilemmas. The urgency to outpace global competitors like China has fueled a narrative that Silicon Valley holds the key to a more liberal world order. Yet, Hao observes that this race has led to a decrease in liberal values, with Silicon Valley emerging largely unscathed. While many celebrate the productivity gains offered by AI tools like ChatGPT, questions remain about the true impact of these technologies. The dual nature of OpenAI as both a non-profit and for-profit entity complicates its mission and raises concerns about blurring the lines between benefiting humanity and pursuing profit. As evidence mounts that AI developments may cause real harm, Hao warns against becoming so entrenched in a belief system that one loses sight of reality. 'There’s something really dangerous and dark about that,' she cautions, urging for a more balanced approach to AI development that considers both innovation and ethical responsibility.