Uoči poremećaja, diskontinuitet u zoru?

Columnist George Will once warned, “The future has a way of arriving unannounced.”

On November 30, 2022, the future did exactly that when the AI genie left the bottle and artificial intelligence democratized. That day, ChatGPT — a large language model (LLM) — was released to the public for free. Within two months, the platform had 100 million users, and AI competitors turned on their afterburners, developing their own versions to keep pace.

Some tech gurus feel ecstatic, while others sound the alarm. The media has broadcasted headlines from around the world. From reducing the hundreds of hours it took a company to find and schedule suppliers to seconds, to writing 10,000 celebrity profiles a month compared to just one, ChatGPT is already transforming society. Other examples abound and more emerge each day: a decent college essay written in 10 seconds; a professor’s course syllabus, class assignments, and grading criteria generated in seconds; etc.

OGLAS

The siloes of society — from government to healthcare to researchers of all ilk — are processing this technological advancement in their own way. Teachers must figure out how to prevent students from using ChatGPT to cheat, while the NIH is assessing what it means for drug development. Artists argue AI training uses their works without compensation, and there’s debate about AI creating new music in deceased singers’ voices.

In effort to provide some guidance, the White House recently issued principles for AI safety, nondiscrimination, equity and privacy. Just this past week, G7 leaders called for “guardrails” around further development of AI — but the technology is not slowing down.

We must accept we are on the cusp of massive societal change that will transcend these siloes, and break their current frameworks for understanding and operating in our world.

Why it is different this time.

OGLAS

The new AI systems are fundamentally different than past waves of automation. Past advancements primarily affected physical or routine tasks that could be codified, written down step-by-step, or programmed into a computer. In contrast, these new systems take on non-routine, creative tasks. Based on examples from their training, they infer what to do in new tasks, thus performing without explicit instructions. As it scales, this general-purpose technology will infiltrate every aspect of the economy and society.

The printing press made knowledge once held only by a few available to the masses, and paved the way for the Age of Enlightenment and our modern world and science. Yet recently, Henry Kissinger, former U.S. Secretary of State, Eric Schmidt, former Google
GOOG
Chairman and CEO, and Daniel Huttenlocher, Dean of MIT’s Schwarzman College of Computing, observed a key difference: the printing press distributed human thought and information, but AI distills it. A decade ago, the mantra was “we’re swimming in sensors and drowning in data.” Suddenly, individuals can query the collective recorded knowledge and experiences of all humanity, and receive a sorted, synthesis in seconds. Everyone is on the verge of being able to access expert systems for whatever they need.

A giant leap in productivity.

Researchers have also started probing the productivity potential of new generative AI systems in work settings. For mid-level writing tasks typical of marketers, consultants, data analysts, human resource professionals and managers, task time was cut by 37% when using AI technology compared to the control group, according to a recent MIT study.^[I] Other studies observed a generative AI-based conversational assistant at scale among thousands of customer support agents. Productivity increased by 14% on average, with the biggest gains among less experienced, lower skilled workers.^[Ii] Further, software developers were asked to use an AI tool to perform a typical task, leading to a 56% faster completion time than the control group.^[Iii]

OGLAS

Imagine productivity gains like this across the economy. About 80% of the U.S. workforce is expected to have at least 10% of their work tasks affected by LLMs, and 19% of workers may see at least half or more of their tasks impacted when AI models combine with complementary software and tools.^iv] Businesses and other organizations cannot ignore these dramatic productivity gains, lest they suffer a competitive disadvantage.

China predicted anticipated major economic and societal discontinuity ahead.

Over the past few decades, the digital revolution unleashed a whirlwind of creative-destruction — thousands of new firms emerged, while others died; entire industries were transformed; new ways of doing business swept the globe; and the ways in which society communicates and socializes were repatterned. Now, imagine that whirlwind on steroids, a Category 5 hurricane sweeping the planet on a compressed timeline. That’s the future of AI.

China knew this day would come, as discussed in the Council’s 2018 Clarion Call. China’s 2017 Next Generation Artificial Intelligence Plan recognized the looming profound changes to human society and life. Their plan noted the potential transformation of employment structures, impact on legal and social theories, and far-reaching effects on the management of government, economic security and social stability. China’s plan was and remains breathtaking in scope — a vision for deploying AI in all forms, from transportation and manufacturing to agriculture, logistics, elder care, and more — in short, within all constructs of society. They’re ready to start implementing those plans now.

OGLAS

Who will rule the new AI world?

In 2017, Vladimir Putin recognized the power of the AI revolution, saying, “The one who becomes the leader in this sphere will be the ruler of the world.”

Control of AI advancement will depend on having both AI technology and massive amounts of data to train it — a major reason why China views data as important a resource as land, labor or capital. Currently, the U.S. and China are frontrunners in the AI race. Both countries account for half the world’s hyperscale data centers, and 94% of all funding for AI start-ups in the last five years, according to UNCTAD’s Izvještaj o digitalnoj ekonomiji 2021. The U.S. also has more than 55% of the world’s AI researchers, and 18 of the top 25 institutions for top-tier AI research.

We need to start asking the right questions.

The LLM-based tools will inevitably go global at an ever-increasing and accelerating pace. If the United States adopts these tools and deals with the ramifications of their use now, then we could have a large advantage. Yet, as our leaders focus on erecting “guardrails,” they seem to miss the metaphorical 18-wheeler barreling down the highway head-on.

OGLAS

As we enter uncharted territory without a playbook, we don’t have many answers, but we can start asking the right questions. Questions such as, how will we deal with the fallout from massive creative destruction? Who or what will benefit from large productivity dividends? How will we manage future waves of new discoveries and scientific advancements? How will work and jobs evolve? These are just a few of the many more questions we must ask. Ultimately, we need to know what kind of new world is at hand.

Every individual and American leader across institutions must recognize that humanity has reached a profound moment. Waves
WAVES
of change are coming at us rapidly. We better become quick and effective at surfing them.