Artificial Intelligence will shape the future of humanity, but who is shaping AI today?

1. The Game-Changer: Deep Neural Networks

Artificial intelligence underwent a major evolution with deep neural networks (DNNs) at the forefront. Modeled loosely after the human brain, DNNs employ simulated neurons to learn and perform tasks without human supervision. This leap allows algorithms to independently develop abilities through "deep learning."

DeepMind's AlphaGo demonstrated just how far DNNs could go. In 2014, this AI program defeated professional Go player Fan Hui in five consecutive games—a milestone that stunned the tech world. Its capabilities arose from being trained on thousands of past Go matches, allowing it to develop strategic thinking at a creative level impossible with older AI models.

Even more impressive, DNNs enable AI to perform increasingly complex and diverse tasks. Industries like healthcare, transportation, and global commerce now integrate these systems to diagnose diseases, improve logistics, and forecast trends. The ability to self-teach has catapulted AI beyond its original coding confines.

Examples

  • AlphaGo's groundbreaking victory against world Go champion Lee Sedol in 2016.
  • The integration of DNNs into Netflix's recommendation systems to understand viewing habits.
  • Development of autonomous cars, relying on DNNs to process real-time driving data.

2. The Birth of Superhuman Intelligence

The creation of AlphaGo Zero in 2017 marked a new era. Unlike its predecessor, AlphaGo Zero required no human knowledge to master Go. It played games against itself repeatedly, learning from each match to create its strategic logic far beyond human understanding.

AlphaGo Zero’s Elo rating skyrocketing to over 5,000 signified its superiority not just to human champions but also to the previous AlphaGo version. In just 40 days, it not only mastered all known strategies but also invented new ones, redefining the game itself. This self-teaching leap into "superhuman" reasoning suggests that AI can create solutions human minds might never anticipate.

This development flips the dynamic between humans and AI. Traditionally, humans taught AI systems. Now, the tables are turning as systems not only learn on their own but surpass human expertise entirely. While this showcases AI's immense power, it also raises concerns about control and unpredictability.

Examples

  • AlphaGo Zero achieved a 90% win rate against the original AlphaGo after self-training.
  • Chess AI engines like Stockfish refining their skills by playing millions of automated games.
  • Self-taught advancements in AI-driven protein folding research by DeepMind's AlphaFold.

3. Accelerating Progress in Narrow Intelligence

Artificial Narrow Intelligence (ANI) flourishes today, with AI excelling in focused, specific domains. These tightly scoped programs are already integral to daily life, from spam filters to virtual assistants like Siri and Alexa. They approach or surpass human ability in designated areas but remain limited to single-task operations.

The rapid deployment of ANI systems has been transforming workplaces. For example, algorithms help hospitals prioritize patient care using diagnostic tools, while financial sectors rely on AI to assess loan risks accurately. Each task benefits from ANI's precision and speed, yet its narrow focus confines it to pre-defined problem areas.

As ANI expands into more industries, society faces increasing reliance on its functionality. Autonomous cars, algorithm-driven surgeries, and financial trend analysis are just a few examples that highlight both the promise and dependency on ANI systems in the modern world.

Examples

  • The convenience of voice-activated assistants like Amazon’s Alexa in performing everyday tasks.
  • AI-powered fraud detection systems analyzing millions of transactions across banks.
  • Self-driving car companies such as Waymo using ANI technology in testing vehicles.

4. The Path Toward General and Super Intelligence

Looking ahead, the transition from ANI to Artificial General Intelligence (AGI) looms. Unlike narrow systems, AGI will be capable of performing diverse tasks akin to human intelligence—including reasoning, decision-making, and adaptability. By the 2040s, AGI may emerge, quickly advancing into Artificial Super Intelligence (ASI), which will surpass human intellect by trillions of times.

This evolution holds the potential to unlock breakthroughs in science, education, and sustainability. However, the risks remain daunting. Systems that continuously self-improve at exponential speeds may spiral beyond human understanding, raising concerns about whether humanity can still intervene.

The trajectory of AI organizations will heavily determine the outcomes. With AGI on the horizon, the race among nations ensures competition to achieve breakthroughs first. This swift innovation leaves little room for governments or industries to plan long-term safeguards.

Examples

  • Predictions of AGI by leading experts like Ray Kurzweil by the mid-2040s.
  • AI projects using reinforcement learning to teach robots complex human behaviors.
  • Medical AI systems diagnosing rare diseases better than seasoned doctors.

5. National Ambitions Shape AI’s Future

Both the US and China dominate AI development, but their paths diverge. In the United States, AI innovation thrives under free-market capitalism, where profit often eclipses ethical considerations. Companies like Google, Microsoft, and Amazon race to release systems without thoroughly examining societal impacts.

China, on the other hand, promotes centralized control. Its government’s $2 billion AI research park and social credit score projects reveal its dual priorities: monitoring citizens and achieving global power. By 2030, China intends to lead the world in AI, cementing its dominance through robotics, surveillance, and scientific development.

These differing strategies—profit-driven innovation in the US and state-fueled AI in China—will define the future of global power structures, accelerating technological advancement while amplifying risks of misuse.

Examples

  • Alibaba's AI systems speeding up logistics in China’s e-commerce.
  • Facebook’s hasty rollouts posing global concerns, as seen in the Cambridge Analytica scandal.
  • The Chinese government funding compulsory AI courses in high schools.

6. The Risks of Global AI Dependency

As AI penetrates every aspect of life, humans risk becoming overly dependent on it. From transportation networks to financial systems, even minor glitches could wreak havoc. The looming arrival of advanced nanobots heightens these risks as healthcare may one day rely entirely on AI-controlled microdevices.

The stakes become more alarming when considering the threat of cyberattacks. Imagine a scenario in which an opposing nation hacks into America’s AI systems. Medical backbones, transportation grids, and even private information could be compromised in seconds—potentially leading to catastrophic outcomes.

These dependencies tether society to AI’s performance. Without preemptive measures, systems capable of operating independently may seize control far beyond our grasp, jeopardizing human agency.

Examples

  • Wearable health monitors like Fitbit relying on AI suggestions.
  • Allegations of Russian hacking of US democratic systems illustrate digital vulnerability.
  • The integration of AI systems into national defense could heighten risks of militarized AI hacking.

7. A Call for Ethical and Sustainable AI

AI developers must embrace greater responsibility toward humanity. But the US tech ecosystem, driven by short-term profits and competitive pressures, lacks the bandwidth for ethical reflection. The Facebook privacy scandals highlight this systemic flaw: companies act first, apologize later.

Thorough testing or simulations may provide a safeguard for future AI—but such efforts demand substantial funding and decreased corporate pressure for immediate profit. Without intervention, companies will continue pushing potentially harmful products into the public domain.

A restructuring of priorities—guided by responsible policies—is necessary. Shifting focus from profits to sustainability can ensure AI aligns better with human values and positive societal goals.

Examples

  • US government incentives encouraging safe AI adoption.
  • Apple's oversight efforts in consumer protection and smart tracking.
  • IBM's ethical AI initiative to adapt algorithms responsibly.

8. Cooperation Through Global Alliances

The fragmentation between US and Chinese AI goals jeopardizes the collective advancement of humanity. A global alliance—like the suggested Global Alliance of Intelligence Augmentation (GAIA)—could unite governments, companies, and specialists to pursue conscientious AI expansion.

Acting as an arbiter, such alliances would encourage all nations to adopt responsible AI ethics while improving transparency. If successful, even rival nations such as China might join to avoid being excluded from major technological advancements.

Building meaningful collaboration requires unified frameworks combining leadership and shared research that upholds principles benefitting humanity.

Examples

  • Proposed GAIA frameworks to improve AI international security.
  • EU-led efforts on data privacy protection inspiring global laws.
  • Joint AI forums promoting science diplomacy and cooperation.

9. The Window is Closing Fast

AI will solidify its trajectory in the next two decades. By 2040, AGI systems will transition into autonomous superintelligence, preventing direct control over their capabilities. Delaying ethical oversight today may leave devastating consequences as systems evolve independently.

The pressure to act rests upon governments and global industries to create standardized policies. Funding AI transparency must be prioritized alongside future-oriented education systems equipping younger populations with AI literacy.

Without immediate reform, humanity risks permanently setting AI on an unpredictable, perhaps irretrievable path.

Examples

  • Policies mandating transparency like GDPR in Europe as early steps toward oversight.
  • AI literacy courses in schools preparing a more informed future workforce.
  • Collaboration frameworks similar to nuclear arms treaties to prevent AI misuse.

Takeaways

  1. Advocate for strong domestic laws governing ethical AI practices and responsible corporate behavior.
  2. Establish global collaborations to pool resources and share breakthroughs sustainably.
  3. Integrate AI literacy programs in education systems to empower future leaders and citizens.

Books like The Big Nine