![\"\"](\"https:\/\/world.edu\/wp-content\/uploads\/2023\/06\/AI_Brain_Human_Robot_Hands.jpg\")
\nDoomsaying is an old occupation. Artificial intelligence (AI) is a complex subject. It\u2019s easy to fear what you don\u2019t understand. These three truths go some way towards explaining the oversimplification and dramatisation plaguing discussions about AI.<\/p>\n
Yesterday outlets around the world were plastered with news of yet another\u00a0open letter claiming<\/a>\u00a0AI poses an existential threat to humankind. This letter, published through the nonprofit Center for AI Safety, has been signed by industry figureheads including\u00a0Geoffrey Hinton<\/a>\u00a0and the chief executives of Google DeepMind, Open AI and Anthropic.<\/p>\n However, I\u2019d argue a healthy dose of scepticism is warranted when considering the AI doomsayer narrative. Upon close inspection, we see there are commercial incentives to manufacture fear in the AI space.<\/p>\n And as a researcher of artificial general intelligence (AGI), it seems to me the framing of AI as an existential threat has more in common with 17th-century philosophy than computer science.<\/p>\n When ChatGPT was released late last year, people were delighted, entertained and horrified.<\/p>\n But ChatGPT isn\u2019t a research breakthrough as much as it is a product. The technology it\u2019s based on is several years old. An early version of its underlying model, GPT-3, was released in 2020 with many of the same capabilities. It just wasn\u2019t easily accessible online for everyone to play with.<\/p>\n Back in 2020 and 2021,\u00a0I<\/a>\u00a0and many\u00a0others<\/a>\u00a0wrote papers discussing the capabilities and shortcomings of GPT-3 and similar models \u2013 and the world carried on as always. Forward to today, and ChatGPT has had an incredible impact on society. What changed?<\/p>\n In March, Microsoft researchers\u00a0published a paper<\/a>\u00a0claiming GPT-4 showed \u201csparks of artificial general intelligence\u201d. AGI is the subject of a variety of competing definitions, but for the sake of simplicity can be understood as AI with human-level intelligence.<\/p>\n Some immediately interpreted the Microsoft research as saying GPT-4\u00a0is<\/em>\u00a0an AGI. By the definitions of AGI I\u2019m familiar with, this is certainly not true. Nonetheless, it added to the hype and furore, and it was hard not to get caught up in the panic. Scientists are no more immune to\u00a0group think<\/a>\u00a0than anyone else.<\/p>\n The same day that paper was submitted, The Future of Life Institute\u00a0published an open letter<\/a>\u00a0calling for a six-month pause on training AI models more powerful than GPT-4, to allow everyone to take stock and plan ahead. Some of the AI luminaries who signed it expressed concern that AGI poses an existential threat to humans, and that ChatGPT is too close to AGI for comfort.<\/p>\n Soon after, prominent AI safety researcher Eliezer Yudkowsky \u2013 who has been commenting on the dangers of superintelligent AI\u00a0since well before<\/a>\u00a02020 \u2013 took things a step further.\u00a0He claimed<\/a>\u00a0we were on a path to building a \u201csuperhumanly smart AI\u201d, in which case \u201cthe obvious thing that would happen\u201d is \u201cliterally everyone on Earth will die\u201d. He even suggested countries need to be willing to risk nuclear war to enforce compliance with AI regulation across borders.<\/p>\n One aspect of AI safety research is to address potential dangers AGI might present. It\u2019s a difficult topic to study because there is little agreement on what intelligence is and how it functions, let alone what a superintelligence might entail. As such, researchers must rely as much on speculation and philosophical argument as evidence and mathematical proof.<\/p>\n There are two reasons I\u2019m not concerned by ChatGPT and its\u00a0byproducts<\/a>.<\/p>\n First, it isn\u2019t even close to the sort of artificial superintelligence that might conceivably pose a threat to humankind. The models underpinning it are slow learners that require immense volumes of data to construct anything akin to the versatile concepts humans can concoct from only a few examples. In this sense, it\u2019s not \u201cintelligent\u201d.<\/p>\n Second, many of the more catastrophic AGI scenarios depend on premises I find implausible. For instance, there seems to be a prevailing (but unspoken) assumption that sufficient intelligence amounts to limitless real-world power. If this was true, more scientists would be billionaires.<\/p>\n Cognition, as we understand it in humans, takes place as part of a physical environment (which includes our bodies) \u2013 and this environment imposes limitations. The concept of AI as a \u201csoftware mind\u201d unconstrained by hardware has more in common with 17th-century\u00a0dualism<\/a>\u00a0(the idea that the mind and body are separable) than with contemporary theories of the mind existing as\u00a0part of the physical world<\/a>.<\/p>\n Still, doomsaying is old hat, and the events of the last few years probably haven\u2019t helped. But there may be more to this story than meets the eye.<\/p>\n Among the prominent figures calling for AI regulation, many work for or have ties to incumbent AI companies. This technology is useful, and there is money and power at stake \u2013 so fearmongering presents an opportunity.<\/p>\n Almost everything involved in building ChatGPT has been published in research anyone can access. OpenAI\u2019s competitors can (and have) replicated the process, and it won\u2019t be long before free and open-source alternatives flood the market.<\/p>\n This point was made clearly in a memo\u00a0purportedly leaked<\/a>\u00a0from Google entitled \u201cWe have no moat, and neither does OpenAI\u201d. A moat is jargon for a way to secure your business against competitors.<\/p>\n Yann LeCun, who leads AI research at Meta, says these models should be open since they will become public infrastructure. He and many others are\u00a0unconvinced by the AGI doom<\/a>\u00a0narrative. A NYT article on the debate around whether LLM base models should be closed or open.<\/p>\n Meta argues for openness, starting with the release of LLaMA (for non-commercial use), while OpenAI and Google want to keep things closed and proprietary.<\/p>\n They argue that openness can be\u2026<\/p>\n — Yann LeCun (@ylecun) May 18, 2023<\/a><\/p><\/blockquote>\n Was ChatGPT a \u2018breaththrough\u2019?<\/strong><\/h4>\n
I don\u2019t consider AI an imminent existential threat<\/strong><\/h4>\n
Why the sudden concern?<\/strong><\/h4>\n
\n\n