<\/p>\n
Artificial intelligence entered the public consciousness wrapped in a promise that was hard to ignore. Systems capable of writing programs, analyzing enormous amounts of information, assisting researchers, translating languages, or collaborating on the discovery of new drugs seemed to herald a transformation comparable to the advent of the internet. Every few months, a model more powerful than the previous one would emerge, and leading companies competed to see who could make the fastest progress.<\/p>\n
Now the narrative has changed.<\/p>\n
Dario Amodei, CEO of Anthropic, recently called on the industry to slow down the pace of development of the most advanced systems. Sam Altman of OpenAI and Elon Musk supported the need to proceed with greater caution. The warnings focus particularly on a new generation of agents capable not only of answering questions but also of performing actions with increasing autonomy.<\/p>\n
These concerns are not trivial. For years, researchers have been studying issues such as the use of artificial intelligence for cyberattacks, the development of biological weapons, the mass manipulation of information, and the possibility that highly autonomous systems might carry out actions different from those intended by their creators.<\/p>\n
But there is another side to the debate.<\/p>\n
Aidan Gomez, co-founder of the Canadian company Cohere, has warned that making the most extreme scenarios the focus of the conversation may end up undermining extraordinary opportunities. Artificial intelligence is also being used to research drugs, analyze diseases, develop new materials, optimize energy systems, and accelerate scientific processes that previously could take years of work.<\/p>\n
Halting its development, therefore, would not be a decision without consequences either.<\/p>\n
And that\u2019s where a paradox emerges that\u2019s hard to ignore. The companies that developed the most powerful systems on the planet are now among those warning that those very systems might be advancing too quickly. Companies that compete fiercely for talent, infrastructure, and users suddenly agree that there are limits that perhaps shouldn\u2019t be crossed just yet.<\/p>\n
The debate has even moved beyond the purely scientific realm. Calls to slow down the development of artificial intelligence recently sent shockwaves through the tech markets, while a lawsuit filed in the United States accuses several of the industry\u2019s leading companies of illegally coordinating their efforts to slow the pace of development. The companies will have to respond to these allegations, but the episode demonstrates the extent to which a discussion that once seemed confined to laboratories is now beginning to affect markets, governments, and consumers.<\/p>\n
For those of us who use artificial intelligence on a daily basis, there's also an issue that's much closer to home.<\/p>\n
Every delay in progress can also mean a medical tool that takes longer to arrive, scientific research that takes more time, educational technology that hasn\u2019t been developed yet, or a solution that could vastly increase the productivity of millions of people. At the same time, moving forward without sufficient safeguards could introduce risks that we would only understand once it was too late.<\/p>\n
That's why perhaps the real debate isn't simply whether we should speed up or slow down.<\/p>\n
The question is who sets the limit.<\/p>\n
So far, a huge part of this technology\u2019s future is being decided within a small number of private companies that own the models, the data centers, and much of the knowledge needed to build them. Governments are trying to keep up with them through regulations, while scientists, entrepreneurs, and citizens watch a technology that is evolving much faster than the institutions tasked with overseeing it.<\/p>\n
Artificial intelligence promised to change the world.<\/p>\n
What's surprising is that, just as it's beginning to show how much it could change things, some of those who contributed most to building it are saying that perhaps the time has come to take our foot off the gas.<\/p>\n
<\/p>\n
La inteligencia artificial lleg\u00f3 al p\u00fablico envuelta en una promesa dif\u00edcil de ignorar. Sistemas capaces de escribir programas, analizar enormes cantidades de informaci\u00f3n, ayudar a investigadores, traducir idiomas o colaborar en el descubrimiento de nuevos medicamentos parec\u00edan anunciar una transformaci\u00f3n comparable con la llegada de internet. Cada pocos meses aparec\u00eda un modelo m\u00e1s potente […]<\/p>\n","protected":false},"author":284226809,"featured_media":1359,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"_jetpack_newsletter_access":"","_jetpack_dont_email_post_to_subs":false,"_jetpack_newsletter_tier_id":0,"_jetpack_memberships_contains_paywalled_content":false,"_wpcom_ai_launchpad_first_post":false,"_jetpack_hide_featured_image":false,"_jetpack_featured_image_duplicate_dismissed":0,"_jetpack_feature_clip_id":0,"_jetpack_memberships_contains_paid_content":false,"footnotes":"","jetpack_publicize_message":"","jetpack_publicize_feature_enabled":true,"jetpack_social_post_already_shared":true,"jetpack_social_options":{"image_generator_settings":{"template":"highway","default_image_id":0,"font":"","enabled":false},"version":2},"jetpack_post_was_ever_published":false},"categories":[33205457,3213,33205425,33205424],"tags":[33205602,33205600,33205595,33205599,33205597,33205598,33205473,33205593,33205603,33205592,33205594,33205604,33205596,33205601,33205567],"class_list":["post-1355","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-al-dia","category-ciencia","category-inteligencia-artificial","category-tecnologia","tag-agentes-de-ia","tag-aidan-gomez","tag-anthropic","tag-cohere","tag-dario-amodei","tag-elon-musk","tag-futuro","tag-ia","tag-innovacion","tag-inteligencia-artificial","tag-openai","tag-regulacion","tag-sam-altman","tag-seguridad-de-la-ia","tag-tecnologia"],"acf":{"introduccion":"