A year has passed since OpenAI launched ChatGPT as a “research preview,” a chatbot powered by a large language model (LLM).
LLMs utilize transformer neural network technology, which emerged from a 2017 paper by Google. ChatGPT offered an accessible interface for GPT-3.5, quickly becoming the fastest-growing consumer technology, gaining over a million users within just five days of its release. Today, it boasts hundreds of millions of users, along with numerous similar bots employing various LLMs from different companies, including the newly introduced Amazon Q, targeted at business applications.
These advancements are poised to transform creative and knowledge-driven tasks. An MIT study conducted last summer revealed that ChatGPT reduced the time required for tasks like writing cover letters and cost-benefit analyses by 40%, while output quality, as judged by independent evaluators, increased by 18%.
The potential of this technology draws comparisons to electricity and fire, as AI stands to fundamentally reshape our lives—impacting work, communication, and complex problem-solving, much like electricity revolutionized power and industry.
The Economic Impact of AI
Consulting giant McKinsey predicts that generative AI could contribute over $4 trillion annually to the global economy. Consequently, tech giants like Microsoft and Google are vigorously competing in this burgeoning market.
Since the advent of ChatGPT, discussions surrounding the implications and safety of AI technologies have intensified. Debates have emerged from places as diverse as the U.S. Congress to Bletchley Park—the historic center for British code-breaking during World War II. These discussions broadly categorize into two camps: AI “accelerationists,” who advocate for rapid technological advancement and acknowledge its vast benefits, and “doomers,” who call for caution, highlighting the risks of unregulated AI progress.
These tensions have spurred the first significant regulatory actions on AI. The EU's AI Act has been in development for years, while the U.S. has enacted a comprehensive Executive Order on “Safe, Secure, and Trustworthy Artificial Intelligence,” aiming to balance innovation with necessary oversight.
Countries worldwide are actively developing AI strategies in reaction to the LLM boom. Notably, Russian President Vladimir Putin has announced a new AI development strategy to counter Western dominance, although he is entering a race where the U.S., China, and the U.K. are already several steps ahead. This announcement seems particularly curious since he previously stated that the nation that leads in AI would dominate the world.
OpenAI’s Secretive Project Q*
The past year has been a whirlwind in the AI landscape. Even dramatic events, like OpenAI's board firing CEO Sam Altman, seemed to peak when he returned in less than a week after investor and employee backlash led to a board overhaul.
Now, speculation surrounds OpenAI’s confidential project known as Q* (pronounced “Q-star”). The name “Q” reflects the “Quartermaster,” the inventive figure synonymous with creating high-tech gadgets in James Bond films.
According to Reuters, the OpenAI board received a warning letter from researchers about advancements related to Q shortly before Altman's ousting, citing potential threats to humanity. Some speculate that the board may have been unaware of Q, which could have contributed to Altman's dismissal. However, it seems unlikely given that Chief AI Scientist Ilya Sutskever also served on the board. Supporting this perspective, reports from Platformer clarify that the board had not received any such warning regarding Q*.
The Road to Artificial General Intelligence (AGI)
Rumors surrounding Q* suggest it may represent a new neuro-symbolic architecture—a significant theoretical development—or a refined integration of existing LLMs and techniques aimed at advancing AI capabilities. Despite the notion that effective neuro-symbolic systems remain elusive at scale, such an architecture could allow AI to learn from less data while offering more transparent reasoning.
Companies and academic institutions, including IBM, are investing in this approach, which they consider a pathway to achieving artificial general intelligence (AGI). Though AGI remains a nebulous concept, it generally refers to AI's ability to process information at a human level or even surpass human capabilities, all at machine speed.
The Atlantic reports that Q* is likely not yet at the forefront of neuro-symbolic breakthroughs; however, its market release could represent a significant stride toward AGI. Nvidia CEO Jensen Huang has suggested that AGI could be attained within five years. In contrast, Microsoft President Brad Smith has emphasized a more cautious timeline, stating that achieving such advanced AI within the next year is highly improbable, suggesting it may take many years or even decades.
Anticipating the Future of AI
The developments we've witnessed with ChatGPT and potential advancements like Q* have sparked varying emotions—optimism, caution, regulatory scrutiny, competition, and speculation. These rapid advancements highlight not only technological milestones but also our ongoing quest for understanding and mastering our creations.
As we look ahead, the coming year promises to be as exhilarating and complex as the last. Our progress will depend on how effectively we navigate the challenges of innovation and regulatory frameworks.
About Gary Grossman
Gary Grossman is the Executive Vice President of the technology practice at Edelman and the global lead of the Edelman AI Center of Excellence.
Ranking
