Falling Walls 2023 round-up: geopolitics, AI and exploited PhD students

09 Nov 2023 | News

Two wars hung over this year’s get-together of scientists and politicians in Berlin. But there was still plenty of excitement, and trepidation, about new breakthroughs in AI, solar energy, and transatlantic science

Germany’s science minister Bettina Stark-Watzinger speaking at this year's Falling Walls conference in Berlin. Photo: Falling Walls Foundation

This year’s Falling Walls conference in Berlin, one of Europe’s biggest annual get-togethers of scientists and politicians, felt at times more like a foreign policy summit than a scientific forum.

And so it might, with wars both in the Middle East and Ukraine upending scientists’ work. The keynote speech on the final day was given by Menachem Ben-Sasson, a former Israeli politician and chancellor of the Hebrew University of Jerusalem, in which he asked the audience for patience as Israel pursues its military campaign against Hamas.

“It’s not about revenge,” he told the audience on 9 November. “Be patient with us”.

“We stand on the side of Israel,” Germany’s science minister, Bettina Stark-Watzinger, told delegates afterwards. Antisemitism “should not have room anywhere”, she said.

Ukraine’s science minister, Oksen Lisovyi, had also been set to speak, but was in the end too busy negotiating with the German government over a new support package to attend.

Despite the geopolitical shadow hanging over this year’s conference, there were many discussions – many even hopeful – about the future of science and technology. Here, Science|Business summarises the key debates.

AI in science

Perhaps inevitably, given the current swirl of debate and policymaking around artificial intelligence, the impact of AI on science and how researchers work was top of mind.

Benoit Schillings, chief technology officer at X, Google’s so-called “moonshot factory”, warned that AI models risk giving seductively convincing scientific explanations that turn out to be completely wrong.

“How do you use these tools without being trapped by your own creation?” he asked the audience during a debate.

And yet, Schilling also argued that AI models could be better at proposing radical new scientific theories, because humans are so resistant to updating their paradigms.

Humans had taken centuries to accept that the sun, not the earth, was at the centre of the solar system, he pointed out.

“AI brings us the ability to find solutions that are counter-intuitive,” he said. “Sometimes a human will reject a solution just because it looks strange.”

AI can help scientists keep on top of a ballooning literature, he said. Schillings has created his own neural network, which suggests new papers based on what he has already read.

He also claimed that he uses Bard, a Google chatbot, to generate new technology ideas for what X should work on next. Inevitably, not all work, but some provide a useful starting point.

Automate the boring

Algorithmic bias in science was also a central topic. Alena Buyx, chair of the German Ethics Council, pointed to algorithms that could detect the risk of renal failure in hospital patients much better in men than women, because they were trained largely on data from men. “This is enshrining a specific bias,” she said.

She also pleaded with the developers of AI systems to create systems that automate boring, routine paperwork for scientists or doctors, rather than the core of their jobs - caring for patients, say. “Make sure we keep the things that are really important to us in our human hands,” she said.

Overall, though, Buyx is cheered by the current intense focus on AI regulation and ethics, which suggests that society was not repeating the mistakes it made in giving total free reign to social media companies. “As an ethicist, I’m happy, and ethicists are never happy,” she told the audience.

Is the PhD up to date?

The conference also heard a debate on whether PhDs are up to date – with one university manager admitting that professors had been artificially extending PhD completion times to exploit their students as cheap research labour.  

“We do see professors keeping PhD students on for a very long time and not letting them graduate because they are contributing to their research. We do see that happening and we’ve actually tried to stop it,” said Alice Aiken, vice president for research and innovation at Dalhousie University in Canada.

Dalhousie had built new “financial structures” to encourage PhD graduation, she explained. “You can’t stay as PhD student forever, doing research for next to no money.”

During the discussion on 8 November, one audience member interjected to ask why no PhD students were on the panel.

Still, despite concerns in Canada that too many people are doing PhDs, Dalhousie intends to increase its cohort. “Let people get the education they want to get,” said Aiken.

An excessive focus on publishing papers while doing a PhD meant students lost sight of the social impact of their work, argued Zainab Kidwai, a dentist who helps treat oral cancer in the developing world. She has recently completed a mid-career PhD, developing a low-cost kit to diagnose the disease.

Instead, despite being a good way to learn project and time management, PhDs were too often still seen as a “rite of passage” before an academic post rather than an end in themselves.

In Canada, 80% of PhD students in the arts and social sciences do not go on into academia, said Aiken. It was therefore “incumbent” on universities to train them for life outside academia – but this task shouldn’t necessarily fall to academics.

“You can’t expect someone who is a career academic to train someone to work in another field,” she said.

Fragile solar supply chains

The EU is falling behind the US and China when it comes to encouraging a home-grown solar panel industry, multiple speakers agreed during a debate on 8 December. China dominates the solar supply chain, arguably making European energy and decarbonisation goals at the mercy of geopolitical flux.

“You can do climate change mitigation with Chinese [solar] modules – if you can get them,” warned Rutger Schlatmann, head of the solar energy division at the Helmholtz Centre Berlin.

He and other speakers agreed that Europe urgently needed to reshore a solar industry of its own, after the once-leading German sector was all but wiped out by cheap Chinese alternatives in the 2010s.

This is important, because without a domestic industry to collaborate with, European scientists will struggle to know exactly what to research, argued Schlatmann.

One research priority is to find ways of making panels from common materials, so that access does not become a geopolitical flashpoint, said Seth Marder, director of the Renewable and Sustainable Energy Institute in Colorado. China currently dominates global silicon production, a crucial component in current panels.

The way China currently manufactures panels “uses tremendous amounts of coal,” he also pointed out, cutting into their environmental benefits.

The US, through its Inflation Reduction Act, India, and China were all pouring money into subsidising their solar industries, but the EU’s own Solar Energy Strategy, launched last year, had not yet made enough of a difference, several speakers agreed during a debate on 8 November.

“We need to have more decisive political action,” said Walburga Hemetsberger, chief executive of SolarPower Europe, a lobbying organisation for the industry. EU state aid rules need to be loosened to allow national governments to invest in their own industries, she said.

Still, the outlook for the technology is still very bright, enthused Schlatmann. It’s already the cheapest source of energy available, and yet there is still “enormous” potential to make more efficient panels, he said. With further advances, solar energy could become incredibly abundant, meaning, “you can make fuels, make new materials…at any place on earth,” he said.

Transatlantic science

Geopolitics might be making scientific collaboration across certain borders harder. But the increasingly prohibitive costs of new scientific infrastructure – the Large Hadron Collider, for example – is pulling in the other direction.

The expense of these projects is now so huge that global scientific partnerships of “likeminded” countries are going to become ever more crucial, said Asmeret Asefaw Berhe, director of the Office of Science for the U.S. Department of Energy, during a debate on 8 November on transatlantic science.

Young-Kee Kim, a particle physicist at the University of Chicago who has been involved in several physics mega-projects, agreed. “These are all done by collaborative efforts across the world,” she said. “Trust building is very important.”

Never miss an update from Science|Business:   Newsletter sign-up