A headshot of Steve Eglas - Executive Director @ Stanford Computer Science Department

Steve Eglash is Executive Director of Strategic Research Initiatives in the Computer Science Department at Stanford University.  Steve develops and manages research programs in data science and artificial intelligence, bridging research and industry, including the Stanford Data Science Initiative, Artificial Intelligence Lab and its affiliates program, Secure Internet of Things Project, Stanford AI Lab-Toyota Center, and DAWN Infrastructure for Usable Machine Learning.  These programs provide opportunities for deep engagement between industry and Stanford’s researchers in big data, AI, IoT, security, autonomous vehicles, robotics, and machine learning.

The data science and AI revolution

There’s been much made of the ‘AI revolution’ in the media. Stories range from scaremongering to sheer utopianism. As one of Stanford’s leading senior computer scientists, Steve speaks clearly and soberly about what lies in store for these technologies. Put simply, he believes that “we are experiencing a revolution in data science and artificial intelligence.”

“Virtually every industry and company is affected by the data science and AI revolution,” explains Steve. “Increasingly, companies are recognizing data as a valuable resource. They’re using data science and AI to achieve improved efficiency and operational excellence. Improved products and services—not to mention entirely new products and services—are leading to improved competitive advantages for many companies.

These shifts in the way data is used and managed are truly transformative. It represents an entirely new approach to enterprise and consumer relations. “Driven by more data, more types of data, cheap storage, powerful computers, new algorithms, and new ways for people to interact with data, this revolution is leading to fundamental changes,” Steve argues. “Data mining and machine learning allow us to identify subtle relationships, distinguish causality and correlation, and predict the future. Clustering and segmentation enable us to act on individuals rather than averages.”

“Increasingly, companies are recognizing data as a valuable resource. They’re using data science and AI to achieve improved efficiency and operational excellence.”

Companies have had access to near-infinite consumer data for years now, so what’s changed? The answer lies in the capabilities machine learning offers enterprises to seamlessly draw insights and answers from those huge banks of data. “We are seeing the development of techniques that allow us to search and extract information from unstructured and semi-structured data with the same ease as structured data,” he says. “This is alongside the rise of statistical and probabilistic approaches that are even more accurate than the deterministic approaches they replace.”

This revolution is not purely technical. These changes are accompanied by concurrent shifts in the way businesses structure themselves and, most importantly, strategize. “Most disruptive of all are the entirely new business models upending many industries. Forward-looking companies are now expanding into adjacent and complementary businesses.”

Natural language processing and machine learning are enabling developments in intelligent customer interactions and retailing. Computer vision and machine learning are enabling practical autonomous vehicles and robots. Genomics is enabling personalized medicine.”

Stanford’s AI research

Established in 1963, Stanford’s AI Lab is one of the oldest in the world. Today, they continue to provide world-leading research, with many of their past accomplishments exerting a direct influence on today’s AI systems—so what are their next steps? Steve explains that “the forefront of AI research at Stanford and elsewhere includes much more human-centred and contextual interactions; automated and assistive decision-making; and continuous learning. Meanwhile, The Stanford AI Lab—Toyota Center for AI Research comprises a portfolio of research projects focused on human-centered AI for interactions between humans and all sorts of intelligent machines—including autonomous vehicles and robots.”

This human-centric approach is having a direct impact on the kinds of technologies being developed right now. In turn, it’s influencing how AI is implemented by industry leaders. “Natural language processing and machine learning are enabling developments in intelligent customer interactions and retailing. Computer vision and machine learning are enabling practical autonomous vehicles and robots. Genomics is enabling personalized medicine.”

Obstacles and regulatory issues facing AI

Figures like Elon Musk might not be wrong to call for caution after all—Steve is very clear about the potential issues associated with AI. He implies it’s up to society as a whole to ensure the benefits of the technology are not distributed unevenly. That means better education and training opportunities, data regulation, and a new social contract with the state. “AI systems have the potential to provide great benefits to society, but they can also be used for evil purposes,” he argues. “Even well-intentioned uses can produce serious problems.”

“For example, autonomous systems will lead to some jobs being eliminated. There is a shortage of workers who are capable of developing advanced AI systems such as autonomous vehicles. Much of the data that enables advanced AI systems is closely held by companies that view that data as a source of proprietary competitive advantage. Society needs to provide education and training for all workers and a safety net for those who are unable to transition to new jobs.”

The AI Summit

We’re looking forward to hearing Steve delivering his keynote address, entitled ‘An AI and data science revolution’ at the AI Summit in San Francisco later this month. He believes that “the AI Summit is an extraordinary opportunity for representatives from diverse companies and industries to engage with each other and with leading researchers in applied AI.”