Eric Schmidt (Alphabet Inc. Technical Advisor) – Conversation with Eric Schmidt (Dec 2021)
Chapters
00:00:01 AI in the Future: Global Trends and Geopolitical Shifts
The Cat Experiment and the Discovery of YouTube’s Cats: In 2011, a group at Google conducted an experiment on YouTube using unsupervised learning. They surprisingly discovered that cats were a dominant theme in YouTube’s content. This led to the creation of Google Brain and the development of language models like BERT and Transformers.
AI’s Impact on Google’s Revenue and Search Quality: Google’s use of AI technologies has significantly improved revenue and search quality. AI’s ability to learn from user interactions and improve the quality of answers and ads has been a key factor in this success.
Traditional Teams vs. Machine Learning Models: Google often pitted traditional algorithmic programming teams against machine learning models built on TensorFlow. While the results were often similar, the AI models often found correlations and patterns that humans could not see, leading to significant gains.
Generative Models and Prediction: Generative models, like GANs, can both identify patterns and generate new content. This has led to the next stage of AI, where predictions can be made with greater certainty.
AI’s Impact on Consumer Tech Companies: AI directly improves customer quality and revenue for consumer tech companies. It allows for individual targeting without profiling, leading to increased customer engagement.
Geopolitical Trends Impacting AI’s Long-Term Benefits: Eric Schmidt emphasizes the need to address geopolitical trends that could affect AI’s long-term benefits. These trends include the rise of China, the US-China relationship, and the global distribution of AI capabilities.
China’s Rise as a Technological Powerhouse: China poses a significant competitive challenge to the United States in terms of innovation and technological advancement. China’s unique model features a large pool of engineers, extensive work hours, and a profit-driven approach, enabling them to exploit platforms at scale.
The Bifurcation of the Internet: The internet has transitioned from an optional tool to a fundamental aspect of human activity. This shift has led to the emergence of two distinct information spaces: China’s internet and the rest of the world’s.
Distinct Features of China’s Internet: China’s internet is characterized by strict censorship and surveillance measures, aimed at maintaining control over its citizens. Anonymity is restricted, and speech is heavily monitored, leading to potential prosecution. American companies face limited access to China’s internet market, while Chinese companies are generally prohibited from operating in the United States, with TikTok being a notable exception.
Tech Industry Culture and Its Impact: The tech industry possesses its own unique culture, values, and political leanings, which may differ from those of the broader society. It is important to recognize and respect the diverse perspectives and values that exist beyond the tech industry.
China’s Strategic Approach: China’s internet policies and infrastructure are designed to maintain the country’s power and control over its citizens. The Chinese government employs surveillance, social credit systems, and other mechanisms to achieve this objective. It is likely that China’s strategy will be successful in achieving its intended goals.
00:10:40 China's Global Influence on Internet Regulation and AI Standards
China’s Digital Authoritarianism: China’s approach to internet regulation is distinct from the US-Western consensus, emphasizing regulation and surveillance. The Chinese government is actively blocking VPNs using machine learning technology.
The Impact on Other Countries: Democracies are likely to align with the US-Western approach, while authoritarian or economically dependent countries may adopt the Chinese model. Countries like Australia have taken a tough stand against Chinese interference despite economic dependence on China.
The BRI and China’s Influence: Countries participating in China’s Belt and Road Initiative (BRI) may become clients of the Chinese information space, adopting Chinese rules and surveillance practices.
Germany’s Dilemma: Germany’s economic reliance on China creates a difficult position as the global divide widens. German companies prioritize China as their top international market and supply base.
The Challenge for Western Partners: Leading Western partners will likely remain in the US-Western fold, but face increasing pressure from China. Countries like Hungary and Tunisia face difficult choices due to economic ties with China.
AI and the Battle for Global Norms: The global divide in internet regulation also extends to AI development and standards. China has been leading the creation of global norms on AI, raising concerns about bias, ethics, and privacy.
00:15:27 China's Rise in AI: Challenges and Opportunities for the United States
The AI Fight is Already Underway: Data is not the only important factor, algorithms are also crucial. National research networks are needed to provide resources to universities and startups. The US must strengthen its own AI practices and not rely on adopting others’ practices.
Increasing American AI Capabilities: Allow high-skilled immigration to attract top talent from around the world. Increase research funding and encourage more talent to join government agencies. Improve technical expertise within government to better understand and address AI challenges.
China’s Progress in AI: China is a strong competitor in AI and is making significant advancements. They are leading in areas like electronic commerce, surveillance, and are catching up in AI. China demonstrated a universal AI model similar in size to OpenAI’s GPT-3, highlighting their progress.
Urgency for Action: The US must act now to maintain its global leadership in technology and values. American technology, values, and startups need a strong foundation to succeed globally. The government and private sector need to collaborate to address this national security challenge.
00:19:21 Government Response to Digital Technology
AI and Digital Technology Innovation: Most innovation in digital technologies occurs in private companies, not government labs or universities. Governments should foster an innovative environment that allows tech companies to advance these technologies without premature regulations.
National Security and Coherent Plans: In national security situations, a coherent national plan is essential. Operation Warp Speed, a collaboration between universities, private companies, and the government, demonstrated the potential for rapid innovation during a crisis.
Leadership and Resource Shifting: Leadership at the presidential level is necessary to drive change in government. Shifting resources and changing priorities are crucial to adapt to the rapid pace of AI and digital technology.
Adapting Government Systems: Government systems often struggle with rapid change and are structured around long-term processes. The government needs to find ways to adapt its funding and deployment processes to keep pace with the rapid evolution of AI.
AI’s Impact on Science and Industries: AI has the potential to transform our understanding of science, biology, chemistry, and material science. The next trillion-dollar industries will be in the application of digital technology and AI in these fields.
Urgency and Time Window: The urgency for governments to address these issues is high due to the rapidly evolving nature of AI and digital technologies. Governments need to operate within a short time window to remain competitive and successfully address the challenges and opportunities of the digital age.
00:23:37 AI Technology: Urgent Need for US-China Balance
A New Era of AI-Fueled Cyber Warfare: TikTok’s success highlights China’s growing technological prowess, challenging the U.S.’s dominance. AI-powered cyber attacks pose significant risks, potentially disrupting global stability.
Urgency to Address the AI Gap: The U.S. has limited time to address the growing AI gap with China. Coordinated efforts are needed to bolster research, establish ethical guidelines, and attract talent.
Recommendations for U.S. Government: Increase funding for AI research and create a national research network. Collaborate with allies to promote shared values and standards. Establish clear guidelines and ethics rules for AI development and use.
Addressing the AI Talent Gap: Proposal for a civilian university focused on technical talent. Creation of a reserve corps to engage experts in government projects. Leveraging public support for national security initiatives.
Critical Problems and Mismatches in AI Technology: Potential for silent and sophisticated cyber attacks using AI. Containment of AI technology is challenging due to rapid knowledge dissemination.
AGI and the Risk of Dangerous Systems: The possibility of AGI systems with devastating capabilities, such as curing cancer or triggering mass casualties. Need for international discussions on AGI nonproliferation and responsible use.
Cyber Warfare and the New Balance of Power: Aggressive use of cyber and influence operations by governments, exemplified by Russian interference in 2016. Balancing power and ensuring control over advanced AI systems is a complex challenge.
Need for a New Doctrine for AI: Lack of established doctrine for regulating and controlling advanced AI systems. Risks of proliferation and misuse of powerful AI capabilities. The urgency to develop international agreements and protocols for responsible AI development and use.
00:29:46 Artificial Intelligence: Implications and Analogy
AI’s Broad Applicability: AI has the potential to transform various industries and domains. It can be used across different sectors, unlike other technologies that create new ecosystems. AI’s ability to be foundational and cross-cutting makes it unique.
Analogies for Understanding AI: The impact of AI is comparable to major technological revolutions like electricity and the internet. It is challenging to find an exact analogy due to the constantly evolving nature of technology.
AI’s Impact on Social Norms: AI and ML can significantly influence societal norms and values. This is because AI operates within the information spaces that shape our morals and behaviors.
AI and Child Development: The use of AI toys that can interact with children raises concerns. As the AI toy becomes more sophisticated, it could potentially influence the child’s behavior and values. This presents a unique challenge as AI’s intelligence approaches human levels.
AI and Misinformation: AI systems can exploit human biases, such as recency bias, to spread misinformation. Targeted misinformation campaigns can manipulate people’s emotions and beliefs.
Machine Learning and Synthetic Biology: Machine learning (ML) is revolutionizing fields like synthetic biology, enabling the design and creation of new biological organisms with specific properties. ML is particularly adept at identifying optimal combinations of compounds, as seen in the development of the drug Halicin, a new general-purpose antibiotic discovered through a collaboration between synthetic biologists and computer scientists. ML’s predictive capabilities extend beyond drug discovery, with potential applications in inventory management, healthcare diagnostics, and personalized medicine.
Scientific and Economic Implications of AI: AI is transforming scientific research, particularly in biology, drug discovery, and material science, by enabling digital simulations and reducing the need for costly laboratory experiments. The economic implications of AI in science are significant, as it can lead to faster and more efficient development of new drugs, materials, and technologies, driving economic growth and innovation.
AI’s Impact on Scientific Discovery: AI’s ability to analyze vast amounts of data and identify patterns has led to groundbreaking discoveries in various fields, including new moves in ancient games like Go and Chess. ML algorithms can sift through complex datasets and uncover hidden insights, leading to novel scientific discoveries and advancements.
AI in Healthcare: ML’s predictive capabilities can be applied to healthcare, enabling accurate diagnosis and personalized treatment plans. By analyzing patient data, ML algorithms can identify patterns and correlations that may be missed by human doctors, leading to improved patient outcomes.
00:36:33 Accelerating Scientific Discovery with AI
Scientific Implications: Recent years have witnessed a slowdown in scientific discovery. AI could potentially reignite scientific progress, fostering a renaissance of discovery. AI can be employed to approximate complex functions and generate new data, enabling breakthroughs in fields like climate modeling and animal behavior research. The development of new instruments has often been a catalyst for scientific progress, and AI can overcome the limitations of expensive traditional instruments by leveraging existing data and natural simulations. AI-driven scientific advancements can pave the way for addressing pressing global challenges like climate change.
Economic Implications: AI’s widespread application across industries and scientific domains has the potential to generate trillions of dollars in value over the coming decades. Industries without strict regulations are particularly susceptible to disruption by AI, as new or existing companies can swiftly adopt AI-powered solutions, leading to market dominance and challenges for competitors. AI-driven advancements can create economic value by increasing productivity, optimizing resource allocation, and enhancing decision-making. The integration of AI into industries can lead to the emergence of new business models, products, and services, driving economic growth and innovation. AI’s impact on job markets is complex, with the potential for both job displacement and the creation of new employment opportunities.
AI and Job Displacement: Eric Schmidt acknowledges that AI brings both benefits and challenges, including job displacement and income inequality. The concentration of benefits among a few winners could lead to a decline in high-quality jobs.
AI for Social Impact: Schmidt emphasizes the need to use AI to solve societal problems, such as addressing bias in parole officer decisions. AI can be leveraged to create a better educated, empowered, and higher income workforce.
Cognitive Work Automation: Alex Wang highlights the rapid automation of cognitive tasks by AI, such as coding and drug discovery. This trend contrasts with the slower automation of manual skills like robotics and self-driving.
Speed of AI Adoption: Industries subject to commercial or regulatory pressure are likely to adopt AI more quickly. Regulated industries face barriers to entry and tend to evolve slower, potentially leading to disparities.
The Need for Inclusive AI: Schmidt questions why industries are not offering AI-powered solutions equivalent to Android and iPhone apps. Inclusive AI adoption can address disparities and create a more equitable society.
00:44:38 Impact of AI and Changing Business Models
Current Challenges: The pace of technological change is surpassing our ability to adapt, leading to challenges in various aspects of life. In education, despite vast resources and efforts, there’s a lack of innovation in the underlying science of learning.
AI’s Impact on Business: AI introduces a new dimension of systems that rapidly improve in quality, effectiveness, and functionality, shifting the underlying physics of business.
Decentralization and Centralization: The current concentration of power in countries like China and large U.S. companies may not be the ultimate state. Technologies may oscillate between centralized and decentralized structures.
Empowerment of Individuals: AI-driven assistants may provide individuals with immense power, potentially shifting the balance of power in society. The formation of future dominant companies is uncertain, as specialized assistants may cater to diverse individual needs.
Business Models of the Future: Past dominant business models, such as advertising, enterprise software, and e-commerce, may evolve. Micropayments, sponsorships, and subsidies may play a role, but building a product that is significantly better than the competition remains key.
The Standard for Success: With AI, the standard for business success may shift towards creating products that are 10 times better than existing solutions. Incumbency advantages, such as regulatory capture and brand recognition, demand a higher standard of improvement for new entrants.
00:51:08 Democratizing AI: Challenges and Opportunities
AI’s Role in Modern Industries: AI has the potential to revolutionize various industries by providing new ways of designing and manufacturing products. Digital twinning is an example of AI’s application in industries like manufacturing and automotive, where a virtual replica of a product is created to optimize its design and performance. Many industries still rely on outdated methods and can benefit from incorporating AI and digital technologies.
Focus on Product Quality, Not Monetization: Eric Schmidt emphasizes the importance of prioritizing product development and user acquisition over immediate monetization. A great product will naturally attract customers and lead to revenue opportunities through licensing, sales, and other methods. Blaming revenue issues on the lack of a revenue plan is often a deflection from the need to build a compelling product.
Addressing Grand Challenges in AI: The current focus in AI is shifting from traditional learning models to sophisticated multi-model reasoning systems. Building an infrastructure that enables master’s students to utilize these powerful models is a crucial step towards democratizing AI. The aim is to make AI tools accessible to a broader range of developers with diverse educational backgrounds.
Democratization of AI and Machine Learning: Eric Schmidt advocates for making AI and machine learning accessible to everyone, not just experts with specialized knowledge. Simplifying tools and providing universal programming models can empower developers to solve complex problems without needing a deep understanding of underlying algorithms. Open-source platforms like GitHub facilitate knowledge sharing and accelerate progress in AI development.
Accelerating Scientific Discoveries: To address the perceived slowdown in scientific discoveries, Eric Schmidt suggests building powerful knowledge platforms that enable the next generation of researchers and developers to build upon existing work. Collaboration and sharing of knowledge through open-source platforms can expedite the pace of scientific progress.
Abstract
Innovation, Competition, and the Future of AI: A Global Perspective – Updated Article
The rapid advancement of artificial intelligence (AI) has become a focal point of global competition, particularly between the United States and China, with profound implications for consumer technology, geopolitics, science, and industry. This article delves into key insights and recommendations, exploring the transformative potential of AI, the urgent need for international dialogue and collaboration, the impact of AI on employment and industries, and the strategies of major global players like China and the U.S. in shaping the future of this technology.
—
The AI Race: China vs. U.S.
At the heart of the global AI competition lies the strategic rivalry between China and the U.S. China’s approach, characterized by a strong work ethic, government support, and a unique internet ecosystem, contrasts with the U.S.’s focus on data, algorithms, and research funding. The U.S. needs to bolster its AI capabilities through increased research funding, building a national research network, and attracting global talent. Simultaneously, China’s internet strategy, centered on surveillance and regulation, presents a distinct model influencing other countries, with democracies tending to align with the U.S.-Western approach and authoritarian regimes gravitating towards China’s model.
China’s Strategic Positioning:
China’s strategic focus on building an internet ecosystem that maintains control and power, exemplified by the Great Firewall and social credit systems, highlights its commitment to becoming a global AI leader. However, this approach poses challenges for countries navigating between economic opportunities with China and democratic values.
The Cat Experiment and the Discovery of YouTube’s Cats:
In 2011, a group at Google conducted an experiment on YouTube using unsupervised learning. They surprisingly discovered that cats were a dominant theme in YouTube’s content. This led to the creation of Google Brain and the development of language models like BERT and Transformers.
AI’s Impact on Google’s Revenue and Search Quality:
Google’s use of AI technologies has significantly improved revenue and search quality. AI’s ability to learn from user interactions and improve the quality of answers and ads has been a key factor in this success.
Traditional Teams vs. Machine Learning Models:
Google often pitted traditional algorithmic programming teams against machine learning models built on TensorFlow. While the results were often similar, the AI models often found correlations and patterns that humans could not see, leading to significant gains.
Generative Models and Prediction:
Generative models, like GANs, can both identify patterns and generate new content. This has led to the next stage of AI, where predictions can be made with greater certainty.
AI’s Impact on Consumer Tech Companies:
AI directly improves customer quality and revenue for consumer tech companies. It allows for individual targeting without profiling, leading to increased customer engagement.
—
Global Geopolitical Shifts
AI is reshaping global geopolitics, with major powers like China and the U.S. influencing international relations through technology and data. This shift is evident in the changing dynamics of internet governance, with China advocating for strict control and the U.S. promoting more freedom and diversity.
Eric Schmidt emphasizes the need to address geopolitical trends that could affect AI’s long-term benefits. These trends include the rise of China, the US-China relationship, and the global distribution of AI capabilities.
The Growing Global Divide in Internet Regulation:
China’s approach to internet regulation is distinct from the US-Western consensus, emphasizing regulation and surveillance. The Chinese government is actively blocking VPNs using machine learning technology.
The Impact on Other Countries:
Democracies are likely to align with the US-Western approach, while authoritarian or economically dependent countries may adopt the Chinese model. Countries like Australia have taken a tough stand against Chinese interference despite economic dependence on China.
The BRI and China’s Influence:
Countries participating in China’s Belt and Road Initiative (BRI) may become clients of the Chinese information space, adopting Chinese rules and surveillance practices.
Germany’s Dilemma:
Germany’s economic reliance on China creates a difficult position as the global divide widens. German companies prioritize China as their top international market and supply base.
The Challenge for Western Partners:
Leading Western partners will likely remain in the US-Western fold, but face increasing pressure from China. Countries like Hungary and Tunisia face difficult choices due to economic ties with China.
AI and the Battle for Global Norms:
The global divide in internet regulation also extends to AI development and standards. China has been leading the creation of global norms on AI, raising concerns about bias, ethics, and privacy.
—
AI’s Societal Impact
Beyond its technological and economic implications, AI is transforming societal norms and human interactions. Its potential to influence behavior, amplify misinformation, and automate tasks, including cognitive ones, raises ethical concerns and the need for careful regulation and oversight.
AI and Child Development:
The use of AI toys that can interact with children raises concerns. As the AI toy becomes more sophisticated, it could potentially influence the child’s behavior and values. This presents a unique challenge as AI’s intelligence approaches human levels.
AI and Misinformation:
AI systems can exploit human biases, such as recency bias, to spread misinformation. Targeted misinformation campaigns can manipulate people’s emotions and beliefs.
—
Scientific and Industrial Revolution
AI’s impact on science is revolutionary, enabling rapid advancements in fields like biology, chemistry, and material science. The discovery of new compounds, such as the antibiotic Halicin, exemplifies AI’s potential to drive scientific breakthroughs and economic growth.
Machine Learning and Synthetic Biology:
Machine learning (ML) is revolutionizing fields like synthetic biology, enabling the design and creation of new biological organisms with specific properties. ML is particularly adept at identifying optimal combinations of compounds, as seen in the development of the drug Halicin, a new general-purpose antibiotic discovered through a collaboration between synthetic biologists and computer scientists. ML’s predictive capabilities extend beyond drug discovery, with potential applications in inventory management, healthcare diagnostics, and personalized medicine.
Scientific and Economic Implications of AI:
AI is transforming scientific research, particularly in biology, drug discovery, and material science, by enabling digital simulations and reducing the need for costly laboratory experiments. The economic implications of AI in science are significant, as it can lead to faster and more efficient development of new drugs, materials, and technologies, driving economic growth and innovation.
AI’s Impact on Scientific Discovery:
AI’s ability to analyze vast amounts of data and identify patterns has led to groundbreaking discoveries in various fields, including new moves in ancient games like Go and Chess. ML algorithms can sift through complex datasets and uncover hidden insights, leading to novel scientific discoveries and advancements.
AI in Healthcare:
ML’s predictive capabilities can be applied to healthcare, enabling accurate diagnosis and personalized treatment plans. By analyzing patient data, ML algorithms can identify patterns and correlations that may be missed by human doctors, leading to improved patient outcomes.
—
Challenges and Opportunities in AI
While AI offers immense opportunities, it also poses significant risks, such as cyber attacks and societal imbalances. The development of AI technologies, particularly in unregulated industries, will disrupt current economic structures, potentially leading to job losses and income disparities. There is an urgent need for digital innovation across all industries to remain competitive.
AI and Job Displacement:
Eric Schmidt acknowledges that AI brings both benefits and challenges, including job displacement and income inequality. The concentration of benefits among a few winners could lead to a decline in high-quality jobs.
AI for Social Impact:
Schmidt emphasizes the need to use AI to solve societal problems, such as addressing bias in parole officer decisions. AI can be leveraged to create a better educated, empowered, and higher income workforce.
Cognitive Work Automation:
Alex Wang highlights the rapid automation of cognitive tasks by AI, such as coding and drug discovery. This trend contrasts with the slower automation of manual skills like robotics and self-driving.
Speed of AI Adoption:
Industries subject to commercial or regulatory pressure are likely to adopt AI more quickly. Regulated industries face barriers to entry and tend to evolve slower, potentially leading to disparities.
The Need for Inclusive AI:
Schmidt questions why industries are not offering AI-powered solutions equivalent to Android and iPhone apps. Inclusive AI adoption can address disparities and create a more equitable society.
—
U.S. Strategies and Recommendations
For the U.S. to maintain its competitive edge, it must increase AI research funding, establish a national research network, and develop consistent AI guidelines and ethics rules. Additionally, hiring skilled personnel for government roles is crucial. Collaboration with allies to develop common AI standards and regulations is essential.
The AI Fight is Already Underway:
Data is not the only important factor, algorithms are also crucial. National research networks are needed to provide resources to universities and startups. The US must strengthen its own AI practices and not rely on adopting others’ practices.
Increasing American AI Capabilities:
– Allow high-skilled immigration to attract top talent from around the world.
– Increase research funding and encourage more talent to join government agencies.
– Improve technical expertise within government to better understand and address AI challenges.
China’s Progress in AI:
China is a strong competitor in AI and is making significant advancements. They are leading in areas like electronic commerce, surveillance, and are catching up in AI. China demonstrated a universal AI model similar in size to OpenAI’s GPT-3, highlighting their progress.
Urgency for Action:
The US must act now to maintain its global leadership in technology and values. American technology, values, and startups need a strong foundation to succeed globally. The government and private sector need to collaborate to address this national security challenge.
—
The Need for International Dialogue
To mitigate potential risks and ensure responsible AI development and use, the U.S. and China should engage in international discussions. Establishing a regime to regulate AGI and promote responsible governance is imperative.
—
Conclusion
The race for AI dominance between the U.S. and China underscores the need for swift action, international collaboration, and responsible innovation. The future of AI is not just about technological prowess but also about shaping a world where technology serves humanity’s broader goals and values.
Language models like CHAI-3DP and ChatGPT transform technology into linguistic interfaces, but their accuracy is limited by the reliability of their training data. AI's linguistic capabilities can be enhanced by computation, enabling precise expressions and systematic idea development....
Meta envisions a future where mixed reality, AI, and smart glasses revolutionize how we connect, learn, and experience the world. Quest 3 and Meta AI Studio exemplify this vision, bridging the gap between physical and digital realms with immersive experiences....
AI is rapidly transforming society, offering both opportunities and risks, while its impact on the job market is complex, leading to job losses in some sectors and increased efficiency in others. AI's advanced capabilities and limitations are becoming clearer, necessitating careful evaluation and mitigation of potential risks....
Neural networks draw inspiration from the brain's structure and are trained to recognize patterns by adjusting their numerous trainable parameters. The Transformer architecture led to significant advancements in AI by introducing residual connections and multi-layer perceptrons for complex problem-solving....
AI systems, like corporations, can develop goals misaligned with human intentions, posing challenges in aligning their behaviors with creator objectives. The emergence of hybrid AIs, combining artificial and human intelligence, raises concerns about their potential power and the need for effective control mechanisms....
Mark Zuckerberg prioritizes developing great products over high-priced devices, focusing on an open ecosystem and accessibility to serve billions of users. He envisions a future where AI is integrated into various aspects of life, enhancing interactions and offering novel experiences....
AI is rapidly changing industries and society, offering benefits but also presenting ethical, societal, and professional challenges, especially in fields like medicine, journalism, entertainment, and education, requiring responsible development and ethical considerations. AI has the potential to enhance humanity and address societal issues like global connectivity and education, but it...