Вештачка интелигенција

The Truth About Artificial Intelligence: Present and Future

Summary

Artificial Intelligence (AI) is the technological story of the year 2023. The CEO of Google states that its impact will be “deeper” than “fire or electricity.” McKinsey joyfully reports that “Generative” AI could potentially “add trillions of dollars in value […]

The Truth About Artificial Intelligence: Present and Future

Artificial Intelligence (AI) is the technological story of the year 2023. The CEO of Google states that its impact will be “deeper” than “fire or electricity.” McKinsey joyfully reports that “Generative” AI could potentially “add trillions of dollars in value to the global economy” on its own, while a coalition of industry leaders signs a statement claiming that “reducing the risks of AI-induced extinction should be a global priority.” Amazon has just invested billions of dollars in an AI company, hoping to compete with Microsoft’s investment in OpenAI. AI passes bar exams, helps students learn, and produces viral videos.

In the face of this hype and investment frenzy, I say – enjoy it while it lasts. Because it won’t last.

Firstly, AI is not progressing as rapidly as the speed of announcements and investments might have us believe. All the major breakthroughs reported this year, all the captivating headlines and impressive demonstrations, are the result of decades of slow, dedicated work happening behind the scenes.

The first puzzle piece in the current AI hype wave – ChatGPT – was literally released as a marketing move to create anticipation for GPT-4, a product that was mostly already built. While there have been subsequent updates and competing models, the difference in AI systems available to the wider public in 2022 and today does not accurately represent the true pace of progress. And while some AI industry leaders talk about the imminent achievement of “Artificial General Intelligence” (AGI) – essentially AI that can truly ‘think’ rather than just perform specific tasks – people were saying the same things in the 2000s, and 70s, and 50s. They were wrong then, and chances are they are wrong now – although that is an entirely different topic.

Furthermore, most of the major improvements in AI capabilities over the past decade stem from scale – increasingly large training datasets that create ever larger and more capable AI models. But that’s not a trend that can realistically continue; according to some estimates, nearly all publicly available text and images on the internet have already been scanned and incorporated into AI models, and that process has generated a massive number of lawsuits that will likely constrain similar data collection in the future. Ironically, AI-generated media is now so widespread on the internet that it will be very hard to avoid being included in future mass collection efforts – which poses a problem as training AI systems on AI data quickly diminishes their capabilities. Increasing the number of trained experts is probably not a sustainable solution – underpaid AI trainers upon whom big tech companies rely are already utilizing available automation tools for ‘training’ that is expected to bring a human touch.

AI will, of course, continue to evolve. But making money from it will be harder than you might think. Large and sophisticated AI models are extremely expensive to develop and maintain, and nearly every ambitious AI project is currently spending large amounts of money – being sustained by either significant institutional resources or cash infusions from wealthy investors who believe they are entering the “next big thing.”

To attract users, most AI products are sold at a loss – and while some users may sign up when the money runs out, many will abandon the tools or switch to less sophisticated but open-source alternatives. And AI’s limitations, such as fragility in uncontrolled conditions and a tendency for “hallucinations” – a problem many experts consider unsolvable – mean that its safe and effective use is not as straightforward. Additionally, Google and Microsoft are embedding AI into their widely used office and productivity products – a move that will likely sink a large number of smaller companies offering similar services but remains to be seen whether it will significantly increase the profit margins of their already present corporate software.

But what about all those rosy financial predictions and big business investments? Well, that’s exactly what we would expect to see in a bubble. There is a strong financial incentive for companies of all sizes to hype AI, invest in it, and claim to be “AI leaders” because during the peak of the bubble, it can quickly inflate the price of their stocks. When AI products actually have to start making a profit, expect that to change very quickly. How many beloved blockchain companies from the end of the 2010s have survived, let alone thrived? How many of those who enthusiastically predicted that those companies would “change the world” are now making the same claims about AI?

Of course, AI will change the world much more than blockchain ever could. It is an incredibly powerful set of technologies with applications ranging from revolutionary to everyday. By positioning ourselves – as individuals, as businesses, and as nations – to confront AI as it truly is, rather than as the most passionate and wide-eyed fans present it, we can achieve a great deal of good with AI. The first step is to set foot on solid ground.

FAQ:

What is Artificial Intelligence (AI)?
– Artificial Intelligence is a technology that enables computers and systems to perform tasks that require intelligence, such as image recognition, language translation, decision-making, etc.

What are the advantages of Artificial Intelligence (AI)?
– Artificial Intelligence can bring numerous benefits, including more efficient task-solving, improved productivity, automation of routine jobs, and innovative solutions in many industries.

What are the limitations of Artificial Intelligence (AI)?
– Artificial Intelligence has several limitations, such as sensitivity to unforeseen situations, the need for large amounts of training data, the possibility of errors, and difficulties in making moral or ethical judgments.

How will Artificial Intelligence (AI) impact the future?
– Artificial Intelligence will likely have a significant impact on all areas of life, including business, medicine, education, transportation, art, etc. It is expected to bring about numerous innovations and improvements but will also pose new challenges and raise ethical and privacy-related questions.