A new type of engineering degree that focuses on "engineering in the era of AI" is proposed by Bashir Al-Hashimi, Dean of the Faculty of Engineering at Southampton University.
LONDON — Traditional electronic engineering degrees should be reimagined to embrace AI and machine learning, according to a prominent UK academic.
Bashir Al-Hashimi, dean of engineering at Southampton University, proposes a new “engineering in the era of AI” degree.
Speaking at the recent TechWorks Summit, a gathering of representatives from the UK “deep tech” industries, Al-Hashimi said UK engineering departments are losing candidates to computer science. This is partially due to increased goverment investment in computing and coding instruction over the last five years. There has not been a similar investment in engineering subjects, he pointed out.
If the UK is to meet its ambitious goal of investing 2.4 percent of its GDP in R&D by 2027, a further 260,000 students are needed. This corresponds to an extra 20,000 graduate engineers every year to sustain the engineering industry, a mammoth task for a sector that accepted just 13,135 students last year.
“What we have are the ambitions and aspirations of government [combined] with universities that are struggling to produce the graduates needed,” he said.
Figures from UCAS, the Universities and Colleges Admissions Service in the UK, show that the number of UK students accepted in electrical, electronic, civil and mechanical engineering programs totaled 13,135, significantly fewer than the 15,430 students accepted in computer science.
“I think this trend is likely to continue. We are seeing that at Southampton [University] already, and talking to colleagues at other universities, they are seeing exactly the same thing: There are more students wanting to study computer science and fewer wanting to study engineering,” he said. “It’s not surprising, given the extent of the media coverage [around AI], AI tends to be associated with computer science.”
Part of the problem is that as AI and machine learning become popular in all kinds of electronic systems, the curricula of UK universities is not keeping pace.
“Currently UK [universities] do not have a systematic approach to updating their engineering curriculum,” Al-Hashimi said. Along with professional societies, Al-Hashimi is urging colleagues to create incentives that "will allow UK universities to embrace digital engineering in a very coordinated way.”
UK engineering departments are losing potential students to computer science (Image: Annie Spratt)
The solution, he said, is to re-imagine engineering education. Al-Hashimi proposes an engineering degree that embraces machine learning and data science in a more coordinated way in teaching both engineering design and practice.
“Twenty-first century engineering is digital engineering… the convergence of engineering practice with machine learning, data analysis, applied mathematics and statistics. It’s built on engineering principles, but in the context of data and intelligence,” he said.
The interdisciplinary engineering degree would qualifies graduates in design, management and training of AI systems along with future intelligent infrastructure. The approach would update mathematics courses to include data visualisation, identifying variables in data and using them in machine learning algorithms. The curriculum could also include some aspects of psychology such as human-machine collaboration, philosophy related to AI ethics and social science to measure the social impact of this burgeoning technology.
Al-Hashimi also advocates a switch to problem-based learning. Subjects such as engineering are commonly taught using design-based learning, where principles are taught first, then problems posed. In case of failure, the result is corrected.
In problem-based learning, knowldge is applied. But it’s expected that there will be gaps in students' knowledge. They are then taught the skills to find the answers needed to solve their problem.
“I personally think what we call design-based learning… this type of teaching may have come to an end,” he said. “There is some foundation in medical education for problem-based learning, where they look at open-ended problems… I feel we need to move away from the design-based approach of learning, to problem-based learning in the era of AI.”
Stronger partnerships between the industry and academia are also essential to ensure the relevance of the new courses to new industries such as AI system design, Al-Hashimi said.
“Professional bodies in particular have a really important role to play," he added. "They accredit programs at universities; they should be looking at the curriculum and saying, 'Is this what produces graduates who are fit for purpose, to enable the UK to be leading in digital engineering?'"
"Sadly, they’re not there yet."