Importance of AI in Business:

Eric Schmidt, the executive chairman of Alphabet, emphasizes the critical role of artificial intelligence (AI) in businesses dealing with large volumes of data. He notes that AI is not just big but essential for maintaining a competitive edge. According to Schmidt, companies involved in knowledge systems will inevitably integrate AI into their core strategies.

Shift in Alphabet’s Core Strategy:

Schmidt reflects on the evolution of Alphabet’s priorities, mentioning that the company transitioned from a “mobile-first” approach to an “AI-first” approach. This strategic shift underlines the increasing importance of AI in shaping Alphabet’s future.

Key Drivers of Competitive Advantage:

When asked about the key factors contributing to a company’s competitive advantage in the AI space, Schmidt identifies two elements: human capital and data. He notes that the field of AI is complex and not fully understood even by scientists. This complexity makes specialized skills, like those possessed by PhDs in mathematics, highly valuable.

Talent Shortage and High Demand:

Schmidt highlights that there is an unprecedented demand for highly skilled mathematicians and data scientists, especially fresh graduates from leading institutions like Carnegie Mellon. These individuals are receiving multimillion-dollar packages, indicating their extreme value in the marketplace.

Role of Training Data:

He further explains that large datasets are crucial for training AI algorithms effectively. Challenges arise when attempting to apply AI to rare occurrences like terrorism due to insufficient data. In contrast, abundant data in areas like traffic management or biology makes AI application more effective.

AI in Healthcare:

Schmidt also touches on the promise of AI in healthcare, emphasizing that humans are biologically similar, providing a rich set of training data. This abundance of data has the potential to lead to significant advancements in global healthcare.

National Security Perspective:

Though not elaborated upon, it’s worth noting that Schmidt’s role as chair of the Defense Innovation Advisory Board implies his insight into the relevance of AI in national security, adding an extra layer of authority to his observations.

Defense Innovation Board’s Role:

Eric Schmidt discussed his experience as the chairman of the Defense Innovation Board. The board’s objective was to modernize the military by aligning it with current technological advancements. They made ten public recommendations aimed at fostering innovation within the military.

State of AI in the Military:

Schmidt noted that the military is not leading in AI and stressed the importance of AI-human collaboration. The traditional use of soldiers for monotonous tasks like surveillance is inefficient when compared to machine capabilities. Computers, especially those with AI capabilities, are more reliable for long-duration tasks and can alert humans only when necessary.

Project Maven & Open Source Data:

Project Maven was cited as a positive example of the military’s foray into AI. It’s an intelligence application that makes recommendations and is trained on open-source data, which is crucial because AI requires large volumes of data for training. Open source data offers a larger set for machine learning compared to classified data.

Challenges in AI Adoption:

Schmidt identified procurement cycles and a lack of technical specification for AI in contracts as major obstacles. The military largely works with contractors who are good at hardware but not software. Hence, despite the military leadership wanting to incorporate AI, they face structural issues.

Talent and Resources:

The Defense Department has the financial means for large, multi-year projects, offering an opportunity for better quality software development. Schmidt argues that the military should make use of softer contracting mechanisms to attract AI talent, given that military salaries may not be competitive in the tech market.

Cultural Shifts Needed:

Schmidt indicated that the adoption rate of new technologies, like cloud computing, within the DoD is slower than desirable. He suggested that the military embrace a “fail fast” Silicon Valley culture of innovation to speed up implementation.

Leadership Perception:

Schmidt concludes that while senior military leadership understands the need for AI, collective action remains a hurdle. The lengthy planning procedures inhibit swift advancements, but there’s an overall recognition that AI must be a priority for the military to remain effective.

China’s AI Ambitions:

Eric Schmidt discusses China’s clear strategy for AI, outlined by their government. By 2020, China aims to catch up to the U.S.; by 2025, they intend to surpass, and by 2030, they plan to dominate AI industries. Schmidt notes that this is a government-led initiative, emphasizing China’s focus and ambition in this area.

Chinese Talent and Innovation:

Schmidt also talks about the talent coming out of China, particularly in the AI field. He refers to China’s top computer science experts and notes that they are very capable. His experiences, like the AlphaGo match, confirm this assessment. Additionally, he cites the performance of Asians, and particularly the Chinese, in global coding contests as further evidence of their capabilities.

Strategic Implications:

According to Schmidt, the U.S. lacks a national AI strategy similar to China’s. He suggests that it’s crucial for the U.S. to focus on the technology, either out of a sense of leadership or necessity. He draws a historical parallel with Sputnik and the subsequent development of NASA, suggesting that the U.S. is at a similar inflection point regarding AI.

Steps for U.S. Competitiveness:

For the U.S. to remain competitive, Schmidt argues that the country needs a national strategy and greater investment in research. He also suggests that immigration policy should be flexible to attract global talent in the field of AI, pointing out that some of the best minds are in countries that the U.S. doesn’t currently allow easy access to.

Military Applications and Safety:

Schmidt also touches on the use of AI in the military, suggesting that AI can make soldiers more efficient but should always have a human in the loop for ethical reasons. He addresses concerns about system failure and hacking by saying that these are “straightforward engineering problems” that can be solved.

Private-Public Partnerships:

Schmidt advocates for partnerships between the military and universities for unclassified, open-source research. He argues that such partnerships benefit the country, the military, and the academic community.

Internal Monitoring of AI:

Finally, the discussion also involves Google’s internal machine learning monitoring systems, suggesting that similar systems could benefit academia and defense. Schmidt agrees and calls for partnerships to share this knowledge.

The transcript suggests that both countries have their strengths and strategies in AI, but the U.S. may need to take more focused action to maintain its competitive edge.

Probable Parity in AI:

The primary speaker, Eric Schmidt, assesses that the U.S. and China will likely be at a similar level in terms of AI development in the next five years. While he doesn’t anticipate China surpassing the U.S. within that time, he notes China’s rate of improvement is “impressively good.”

Self-imposed Constraints:

Schmidt indicates that the U.S. seems to be competing in AI with “one hand behind our back.” Rather than adopting policies similar to China, he advocates that the U.S. should focus on what it excels at, like funding basic research. He criticizes reductions in federal research budgets as a step in the wrong direction.

Private Sector and Education:

On a positive note, Schmidt highlights that the explosion in AI and tech opportunities is largely being funded by the private sector. He points out that computer science is now the most popular major in leading universities, outpacing traditional majors like economics. This trend suggests a strong pipeline of talent for future advancements in AI.

Optimism and Concerns:

Lastly, Schmidt is optimistic about the quality of young scientists and inventors entering the field, stating they may even surpass the older generation in terms of talent and innovation. However, he expresses concern that the U.S. is not framing the challenge of AI leadership correctly, implying a need for a strategic national approach.

AI as a People Problem:

Eric Schmidt responds to a question about the “people problem” in adopting AI, emphasizing that organizational culture and behavior are critical. The question underscores the importance of adapting to AI not just technologically but also culturally, particularly in longstanding institutions like corporations and the government.

Learning from Nuclear Research Models:

Schmidt references the U.S. government’s establishment of specialized labs during the nuclear age as an example of how to approach AI’s unique challenges. In the past, institutions like Sandia were created to tackle complex physics problems, staffed by experts in the field. Schmidt suggests a similar approach for AI, recommending the creation of specialized AI institutes.

AI’s Business Impact:

He also predicts that every major public corporation will rely on AI specialists, either in-house or through suppliers. Schmidt provides a compelling example: a small team of AI experts could potentially increase a large business’s profits by 5% through pattern recognition alone.

Case Study: AI in Energy Efficiency:

Schmidt recounts how DeepMind’s reinforcement learning algorithms improved the energy efficiency of their data centers by 15%, surpassing even the best engineers in the field. He emphasizes the magnitude of such improvements, particularly for sectors like oil and gas and transportation, where such energy efficiency gains could represent significant profit increases.

AI’s Transformative Power:

In summary, Schmidt underscores the transformative potential of AI, not only in technological dimensions but also in affecting organizational culture, efficiency, and profitability. He advocates for the U.S. to develop specialized institutions for AI, similar to what was done for nuclear physics, to harness this technology effectively.

Regulation & Tech Companies:

Eric Schmidt addresses the skepticism within the tech sector about working with the federal government, especially on AI and national security issues. He argues against “premature regulation,” suggesting that defining what to regulate at this stage could be difficult and counterproductive.

AI Arms Race:

Schmidt acknowledges the concern about a potential AI arms race, particularly with China. However, he disputes the automatic assumption that competition in AI capabilities will necessarily lead to an arms race. He also mentions that China is developing AI for both commercial and defense purposes.

Cybersecurity Concerns:

On the issue of cybersecurity, Schmidt argues for enhancing domestic cyber systems before contemplating an arms race. He criticizes the government’s outdated Windows 95 systems as highly vulnerable and suggests that a focus on improving cybersecurity could alleviate some concerns.

Technology Proliferation:

When asked about the potential need to protect certain technologies from being exported, Schmidt questions the effectiveness of such measures. He points out that the industry moves quickly, and there is no “fissile material” in AI that can be easily restricted.

Superintelligence Debate:

Schmidt counters Elon Musk’s pessimistic view of AI by stating that the AI systems being developed are good at imprecise pattern matching rather than possessing ‘superintelligence.’ He believes AI will make people more productive and less prone to conflict, contrary to Musk’s warnings.

DeepMind and Talent Pool:

Schmidt mentions that if anyone is going to invent general intelligence, it’s likely to be Alphabet’s DeepMind subsidiary in London. He attributes this to the brilliant talent pool available there, particularly from Cambridge, and expresses concerns about potential Brexit-related restrictions on talent mobility.

Overall, Eric Schmidt’s perspective is one that advocates for caution in regulatory measures and promotes domestic technological improvements. He also encourages a nuanced dialogue about the implications of AI in both national security and societal contexts.

AI Cyber Power:

Richard Harkner introduces the notion of “AI cyber power” distribution, focusing on the U.S. and China. He raises questions about Alphabet’s role as a strategic actor. He suggests shifting from a partnership model to an alignment model, wherein Alphabet has interests that may align differently with nations.

Alphabet Under U.S. Regulations:

Eric Schmidt clarifies that while Alphabet is a significant player in the AI arena, the company operates under U.S. law and regulations. He also indicates that Facebook and a Chinese company are other major actors. He envisions a future where multiple companies will amass data and expertise to influence the AI landscape.

Corporate Contribution to AI Research:

Jeff Klune discusses the investment Google makes in research, paying researchers to work within the company. Schmidt argues that for an AI-first company, investing in research is crucial. He extends this argument to the military, stating that internal expertise is invaluable for national strategy.

Military and Tech Collaboration:

Questions arise about the relationship between Silicon Valley and the Department of Defense. Schmidt acknowledges the historical hesitancy in the tech community towards military collaborations. He anticipates that new companies aligned with military missions will emerge and leverage AI technologies for national security.

Investment in Small Tech:

Schmidt cites the example of DIUx, which invests in small tech companies that align with military objectives. He suggests that this is a more likely scenario for the future, where specialized companies contribute to national security.

Eric Schmidt’s responses offer a nuanced look at how technology companies like Alphabet are navigating their roles in the evolving landscape of AI, corporate interests, and national security.