Microsoft Co-founder Bill Gates: AI is Expanding at a Pace Beyond Regulation, the World May First Face Some Negative Consequences
Microsoft co-founder Bill Gates warned that artificial intelligence is expanding at a pace that regulation cannot keep up with, and the world may first face some negative consequences before rules catch up. He stated that currently, there is no adequate preparation from governments, experts, and communities for the transition, nor is there a plan to mitigate the entry into the AI era.
In a nearly 6,000-word article on his personal website, he outlined three risks: a significant number of jobs will permanently disappear, with middle-class jobs being the first affected; AI will lower the skill threshold for committing fraud, deepfakes, cyberattacks, and even biological threats; and companion products may weaken children's social skills and critical thinking. He also noted that if left unchecked, systems may be designed according to the needs of the wealthiest groups.
Gates mentioned that he has opposed a simple halt for the past three years, but his stance has changed. Triggers include the crossing of boundaries by proxy programming tools into coding and cyberattack capabilities, with the industry having almost no corresponding responses. He told The New York Times that there is private concern within the industry, but publicly, there are reminders not to speak out to avoid affecting the next round of trillion-dollar funding. He stated that voluntary reviews lack clear thresholds and penalties, refusing to hand over the most dangerous tools to industry self-regulation.
Policy proposals include taxing tokens and robots to curb replacements, designating certain jobs as human-only, mandating reviews for systems that can be used to create new molecules, and promoting international standards similar to nuclear safety and aviation regulation. He has communicated with the U.S. presidential team and plans to seek a meeting with the Chinese president within the year. The Gates Foundation also announced plans to invest at least $1 billion over the next two years to expand the use and assessment of AI in impoverished areas.
He emphasized that geopolitical competition and economic incentives will continue to push countries and companies into a full-speed race, so even if he is willing to support a credible global slowdown plan, he does not believe it can be implemented. Customer service, software, and legal assistance will be cut first, while the construction and hotel industries may increasingly turn to robots within the next decade. He worries that once a backlash occurs, even the benefits in science and medicine will be swept away.
The market mechanism is narrative-driven rather than based on weekly pricing events. Sellers are laboratories and cloud vendors still financing based on computing power and model capabilities, while buyers are governments and foundations that need to externalize risks to taxpayers. Beneficiaries are leading companies that have secured enterprise distribution and data center orders, while those under pressure are entry-level white-collar workers and regulatory agencies that have not yet established review capabilities. Funding continues to flow into training and reasoning, while governance funding remains primarily through initiatives and charitable grants, with the two speeds being asymmetrical.
Source: Public Information
ABAB AI Insight
Gates' recent comments mark a shift from "innovation will fix itself" to "innovation may have net negatives." He shaped the default optimism of the personal computer era, and in the 2010s, the foundation framed technology as a global health tool. In 2023, he still described the employment impact as manageable turbulence; by 2026, he changed his tune, as proxy systems have begun to bridge tools, code writing, and real network access, while the pause line drawn by the industry at that time triggered no mandatory actions.
Resource mobilization is divided into two paths. One is his call for taxing tokens that replace human labor and designating human-only jobs, shifting replacement costs from corporate profit statements back to public budgets; the other is the foundation's separate allocation of AI accessibility within its ten-year, billion-dollar scale to prevent products in a regulatory vacuum from only serving high-income markets. He is also engaging with U.S. and Chinese leaders, acknowledging that bilateral competition will not end, and unilateral slowdown equates to ceding computing power hegemony, thus focusing on establishing common standards for detecting biological and cyber capabilities.
The comparison point is his experience with software antitrust and the subsequent lag in social media regulation: rules were formed when Microsoft faced threats of breakup, while Facebook and short videos proliferated before legislation. OpenAI, Anthropic, and Google DeepMind are raising trillions for infrastructure while discussing voluntary safety frameworks. The industry is in an expansion phase: thresholds have been crossed, yet licensing rights have not shifted from companies to the state.
Structural changes are a result of regulatory lag leading to a transfer of pricing power. Model companies define facts through product speed, while governments acknowledge facts through post-facto taxes and bans. The mechanism is: financing punishes executives who openly discuss negatives, while geopolitical pressures punish countries that unilaterally slow down, thus society first pays the tuition through unemployment, fraud, and youth dependency, before writing that tuition into law. Gates aims to issue this tuition bill in advance rather than letting the market settle it on its own.
ABAB News · Cognitive Laws
Voluntary safety standards often die before the next round of financing
When regulation cannot keep up, society first acts as a testing ground
Those discussing slowdown often have already extracted profits from acceleration