Pramod Khargonekar, distinguished professor of electrical engineering and computer science and vice chancellor for research at the University of California, Irvine
Image: UC Irvine
Last year, ZDNET ran a special feature called, “The Intersection of Generative AI and Engineering,” which explored the tremendous potential of generative AI for software development and product development.
Also: The best free AI courses in 2024
This intersection between AI and traditional engineering is rapidly becoming its own formal discipline called AI Engineering. To explore this, ZDNET had the opportunity to discuss AI Engineering with Pramod Khargonekar, distinguished professor of electrical engineering and computer science and vice chancellor for research at the University of California, Irvine.
He is an expert in control and systems theory, cyber-physical systems, and applications to manufacturing, renewable energy and smart grids, and biomedical engineering. Most recently, he has been working on the confluence of machine learning for control and estimation.
Also: If you want a programming job, be sure to learn these three languages
Khargonekar most recently was the lead author of the National Science Foundation-funded report by the Engineering Research Visioning Alliance (ERVA), entitled “AI Engineering: A Strategic Research Framework to Benefit Society.” The report states that AI Engineering is “A generational opportunity to supercharge engineering for the benefit of society through enhancements to national competitiveness, national security, and overall economic growth.”
So, with that, let’s dive into AI Engineering with Professor Khargonekar.
ZDNET: Can you provide an overview of AI Engineering and its significance in the current technological landscape?
Pramod Khargonekar: AI Engineering is a nascent research direction arising from the convergence and synthesis of AI and engineering. It leverages the traditional strengths of engineering disciplines (ensuring safety, reliability, efficiency, sustainability, and the human-technology interface) with breakthrough developments in the AI field.
A recent report by the Engineering Research Visioning Alliance (ERVA), an initiative funded by the U.S. National Science Foundation (NSF), on which I was the lead author, explains how AI Engineering will be bidirectional and reciprocal. It evokes a future vision in which an engineering approach makes for better AI while AI makes for better-engineered systems.
Also: How I used ChatGPT to scan 170k lines of code in seconds and save me hours of detective work
AI Engineering is based on the firm commitment of engineering processes and culture to ethics of safety, health, and public welfare. Its significance lies in conceptualizing a generational opportunity for research and technological advances in engineering as well as AI.
ZDNET: Can you provide an example of a successful AI Engineering project or initiative?
PK: The use of AI in advancing semiconductor design is a very promising development that is already having a meaningful impact. Many companies in electronic design automation (EDA) are incorporating AI-driven tools in their products, resulting in significant enhancements in efficiency, customizability, performance, and sustainability of the semiconductor design process.
ZDNET: What are some examples of AI enabling more efficient engineering results?
PK: AI is transforming the way we approach engineering. Advances in autonomous systems, such as self-driving cars and unmanned air vehicles, are being enabled by AI.
In manufacturing, machine learning and AI tools are used to improve product quality, resource efficiency, and cost reductions. AI is playing an increasing role in state-of-the-art robots. AI can also improve engineered systems to improve product performance and mitigate rare events of high consequence.
Also: Want to work in AI? How to pivot your career in 5 steps
Examples include minimizing drug side effects, mitigating software security flaws, preventing bridge collapses, averting seismic-induced building damage, and stopping chemical plant failures.
These applications show how AI impacts the cost, performance, efficiency, customizability, and sustainability of engineered products and systems. This leads to significant enhancements to the productivity and capabilities of engineers across all disciplines, from practicing engineers and engineering researchers to engineering educators and students.
ZDNET: What challenges do industries face when integrating AI with traditional engineering practices?
PK: Integrating AI with traditional engineering practices presents several challenges. Modern deep learning-based AI tools require massive amounts of high-quality data. This is a significant bottleneck. Engineered systems require very high levels of safety, reliability, and trustworthiness.
Also: Have 10 hours? IBM will train you in AI fundamentals – for free
These are not easy to achieve with the limitations of current AI technologies. Combining and integrating very large numbers of even simple components into a system or engineered product may lead to the emergence of complex behaviors that cannot be easily predicted.
ZDNET: What role do engineers play in the development of AI systems?
PK: Engineers have a crucial role to play in the development of AI systems. The most obvious is the importance of semiconductor chips for AI model training and inference. In applications where AI is integrated into products requiring high levels of safety and reliability, engineers have a critical role in product design, testing, and operation.
In current AI applications, the consequences of errors are either not severe or are being managed by human supervision. For AI to be fully accepted in broader domains of society, safety, reliability, and trustworthiness must increase. Engineers can help achieve these goals.
ZDNET: Can you discuss the importance of multidisciplinary collaboration in advancing AI Engineering?
PK: AI Engineering vision is inherently multidisciplinary. In the engineering for AI pillar, we expect fields such as integrated circuits, thermal and energy sciences, control systems, information theory, and communications theory to work with machine learning and AI to develop more efficient, sustainable, reliable, safe, and trustworthy AI systems.
Also: The end-to-end AI chain emerges – it’s like talking to your company’s top engineer
We also expect machine learning and AI experts to work with those in engineering design, manufacturing, testing, and operations, as well as materials, chemical, energy, environmental, civil, aerospace, and automotive engineers.
In addition to convergence from within their respective engineering disciplines, ensuring the success of AI engineering would also require the collaboration of leaders from government, universities, industry, civil society, and nonprofits.
Strategic alignments among these sectors will energize collaborative efforts and be essential to secure the financial, technological, organizational, and human resources needed to fully realize the AI Engineering vision. This sector convergence approach will facilitate a crucial element of the AI Engineering enterprise: the computing power and generation, collection, and curation of datasets for engineering-specific AI tools.
ZDNET: What specific skills are required for the next generation of experts in AI Engineering?
PK: AI engineers will need to understand complex systems, manage an expanding trove of heterogeneous data, be aware of the limitations of AI techniques, and be fully skilled in the ethics and compliance aspects of AI Engineering.
Also: Want to be a DevOps engineer? Here’s the good, the bad, and the ugly
The latter is increasingly important in maintaining the security and integrity of AI-driven systems.
ZDNET: What are some potential breakthrough developments in AI Engineering for manufacturing?
PK: As more sensors and smart analytics software are integrated into networked industrial products and manufacturing systems, predictive technologies can further learn and autonomously optimize performance and productivity.
Data-centric metrology systems are a critical area for smart semiconductor manufacturing, which can help yield improvement by overcoming inspection and metrology challenges through accelerated data-centric analytics.
Also: Nvidia will train 100,000 California residents on AI in a first-of-its-kind partnership
Newly emerging generative AI tools can enable gathering, understanding, and synthesizing “voice of the customer” quality feedback and user complaints, which today are labor-intensive processes.
In engineering systems, decisions are often made using large knowledge models (including physical-based models, data-centric models, rule-based reasoning, and human experiences).
ZDNET: How do you envision the future of AI Engineering in terms of industry applications?
PK: We envision a future where AI Engineering techniques and expertise will positively impact design, manufacturing, testing, and operation in many industries.
There is great potential for increased efficiency, waste reduction, and increased resilience. There is potential for creative leveraging and reuse of existing knowledge, designs, and processes.
ZDNET: What steps can private industry take to build capacity for AI Engineering?
PK: Private industry is well positioned to encourage and upskill the workforce and learn about current and future machine learning and AI technologies. In partnership with academic institutions, industry can articulate opportunities for education and training needs.
Also: How do AI checkers actually work?
Industry consortia have the opportunity to tackle the cross-cutting need for high-quality data and domain-specific tools.
Finally, there is a major need for computing and data resources like the National AI Research Resource (NAIRR), that are accessible to a much wider community. Industry can work with government to secure funding for investment in such resources.
ZDNET: How can cross-organizational focus on data, design, testing, and operations benefit AI Engineering?
PK: Within an organization, a holistic approach to data, design, testing, and operations is crucial to success. Across the ecosystem, realizing the full potential of AI Engineering requires convergence, coordination, and collaboration of people and organizations from academia, industry, and government.
Also: So long, point-and-click: How generative AI will redefine the user interface
These efforts will need to address difficult challenges in creating and curating datasets. This is incredibly important given the rapid pace of AI innovation and the urgency raised by global competition.
We need to mobilize large-scale financial, technological, human, and organizational resources now, and that will take strong, proactive, coordinated, and collaborative action by leaders working across sectors.
The resulting benefits will accrue to the organizations that are able to position themselves to lead in this rapidly changing environment.
ZDNET: What are the key research directions that need to be established in AI Engineering?
PK: We identified eight Grand Challenges as key research directions. These are:
- Design safe, secure, reliable, and trustworthy AI systems
- Transform manufacturing quality, efficiency, cost, and time-to-market
- Build and operate AI-engineered systems with cradle-to-grave state awareness
- Overcome scaling challenges in engineering
- Construct engineered systems for safe, reliable, and productive human-AI team collaboration
- Mitigate rare event consequences via AI
- Incorporate ethics in all facets of AI Engineering
- Develop engineering domain-specific foundation models
We also recommend dedicated AI Engineering Research Institutes as well as cross-cutting national initiatives to enable the development of the AI Engineering field.
ZDNET: How can AI Engineering contribute to solving complex engineering problems?
PK: Increasingly capable AI tools can transform fundamental disciplines of engineering science. They can also transform major design, manufacturing, and infrastructure engineering endeavors.
These new capabilities will impact the cost, performance, efficiency, customizability, and sustainability of engineered products and systems. They will increase the scope of engineering to address complex societal problems.
Also: The best AI for coding in 2024 (and what not to use)
They will also significantly enhance the productivity and capabilities of engineers across the full spectrum of the discipline: practicing engineers, engineering researchers, engineering educators, and engineering students.
ZDNET: What are the ethical considerations surrounding AI Engineering?
PK: AI Engineering technologies should be designed for augmenting and serving humans. We call for the development of an ethical matrix for AI Engineering.
Such an ethical matrix is envisioned as a practical, pluralistic tool, drawing from traditions that focus on promoting well-being, autonomy, and justice as fairness. It encourages users to examine matters systematically, considering the point of view of each affected group.
ZDNET: How can AI Engineering improve sustainability in various industries?
PK: One example is to bring a sharp focus on reducing energy consumption of data centers, which are central to the development and implementation of current and future AI technologies.
In addition, AI Engineering can create powerful technologies for energy efficiency, renewable electric grids, energy storage, decarbonization of manufacturing cement and metals, and sustainable materials.
ZDNET: How can AI Engineering be used to enhance safety and reliability in engineering projects?
PK: AI Engineering envisions a future in which an engineering approach makes for better AI while AI makes for better-engineered systems. AI Engineering is based on the firm commitment of engineering processes and culture to ethics of safety, health, and public welfare.
The context of safe, secure, reliable, and trustworthy AI systems offers a prime example. AI safety has three distinct but complementary dimensions:
- Assuring a deployed AI system is safe and reliable
- Using an AI system to monitor and improve the safety and reliability of a (potentially non-AI) system/platform, and
- Maximizing safety and trust in collaborative human-AI systems.
AI systems are fast becoming prevalent and influential in society, so ensuring their safety and reliability is critical. A focus on engineering AI safety can help prevent harmful outcomes, mitigate risks, ensure that AI technologies are developed and used responsibly, and help AI systems achieve their full potential.
ZDNET: What impact do you think AI Engineering will have on the future job market?
PK: We think it will impact existing jobs by automating some routine steps and tasks. This will make current workers more efficient and productive.
Also: From AI trainers to ethicists: AI may obsolete some jobs but generate new ones
But a much larger impact will depend on conceptualization and development of new industries and jobs that don’t currently exist.
AI Engineering can help address major human needs such as health and wellness, education, housing, energy, water, food, etc., in the United States and across the world.
ZDNET: How can AI Engineering support innovation in product design and development?
PK: One of the frequently used skill sets in product design and operations of complex engineering systems is exploring new design options, identifying root causes, and tracking solutions for a complex engineering system. This requires time-intensive efforts to recreate issues in lab environments so appropriate solutions may be found.
Newly emerging generative AI tools can enable gathering, understanding, and synthesizing “voice of the customer” quality feedback and user complaints, which today are labor-intensive processes.
Also: The AI scams infiltrating the knitting and crochet world – and why it matters for everyone
Suitably trained, they have the potential to generate new designs in an iterative process led by a design engineer.
ZDNET: What advice would you give to young professionals interested in pursuing a career in AI Engineering?
PK: Much of the academic infrastructure needed for AI Engineering to flourish must be built out by higher education and policy leaders in tandem with private industry. Young professionals interested in engineering should take as many courses as possible related to AI and ensure it remains a focus.
Likewise, those studying AI should also understand how it intersects with engineering. As AI Engineering develops, those with the foresight to understand the connectedness of AI and engineering will be in a great position to advance.
To the future and beyond
AI seems to be a force multiplier across engineering disciplines. Of course, AI also has its limitations. It will be up to the engineers who use and rely on AI to tap into its strengths while compensating for its weaknesses.
What do you think? Are you applying AI to your projects now? Are you looking forward to the new doors AI may open in R&D and product development? Or are you, like me, watching with cautious optimism, but also expecting inevitable failings and foibles along the way? Let us know in the comments below.
You can follow my day-to-day project updates on social media. Be sure to subscribe to my weekly update newsletter, and follow me on Twitter/X at @DavidGewirtz, on Facebook at Facebook.com/DavidGewirtz, on Instagram at Instagram.com/DavidGewirtz, and on YouTube at YouTube.com/DavidGewirtzTV.
