At Rise of AI 2019 in Berlin, founder Fabian Westerheide opens the conference with an overview of the state of AI. He recalls that the event started five years earlier as a small meetup called The Singularity Is Near in a Berlin hackerspace and has grown to about 800 attendees. Drawing on an AI Index study he worked on with Stanford University, he ranks the United States first globally, citing its research universities and large tech companies, and China second, noting a focused national strategy, abundant data and roughly 130 billion US dollars invested in AI. He argues Europe lags behind, lacking a coherent political agenda, large digital companies, enough startups and capital, despite strong research institutions such as DFKI and Max Planck. He describes AI becoming political, citing France's national strategy and a less concrete German one, and becoming big business, with the conference's own exhibition space growing. He argues most companies are not ready for AI because they have not digitized their data, and that deep learning, while not over, has stopped surprising researchers. He closes with open questions on ethics, bias and AGI, and announces the next conference date.
Key takeaways
- Westerheide says Rise of AI grew from a roughly 50-person Berlin meetup five years earlier to about 800 attendees in 2019.
- He ranks the United States first in AI globally and China second, citing about 130 billion US dollars China has invested in AI.
- He argues Europe lacks a coherent AI political agenda and enough startups and capital, despite strong research institutions such as DFKI and Max Planck.
- He argues most companies are not ready for AI because they have not yet digitized their data and processes.
- He cites about 21,000 AI companies founded in the last five years, estimating roughly 10,000 are not purely marketing-driven.
Detailed notes
From a Berlin meetup to an 800-person conference
Westerheide recalls starting the event five years earlier as a meetup called The Singularity Is Near, held in a Berlin hackerspace, after reading a book by that name and finding AI was not yet discussed as a coherent theme. Attendance grew from an initial 50 people with a long waiting list to 170, then 300, then 600 last year, and about 800 this year. He tells the audience that working in AI now, whether running a company or applying it inside one, positions them at the right time in a field short of talent, capital and service providers.
The AI Index: the US, China and a lagging Europe
Building on an earlier study, Westerheide says he worked with Stanford University this year on an AI Index that scores nations on politics, industry, data, researchers, talent, startups and capital. He places the United States first, citing its large tech companies, top universities such as MIT and Stanford, and capital, despite weak federal AI politics. He ranks China second, pointing to a focused national strategy, a strong ecosystem, abundant data, more startups than Europe, and roughly 130 billion US dollars invested in AI, though he says China still needs to hire European researchers. He says Europe lags: no European company ranks among the world's 50 largest digital companies, data protection law makes working with data harder, and there are not enough startups or capital, even though research institutions such as DFKI and Max Planck are strong. He frames Rise of AI itself as an attempt to build a European AI ecosystem and a bridge for Asian and American companies.
AI becomes political and big business
Westerheide says AI only became a political topic in the past year, pointing to France's national AI strategy and a less concrete German government strategy, and argues that once an issue becomes political it tends to stay relevant for years as it trickles down to local governments. He also describes AI becoming big business, noting that the conference's own exhibition space had to expand after turning companies away the previous year, and describes an emerging value chain from chip makers to infrastructure platforms to vertical applications.
"AI is not the answer" and deep learning's plateau
He argues AI is not the answer to every problem, since most companies have not yet digitized their data and processes and need a broader strategy before adding AI on top. On the recurring question of whether deep learning is over, he says it is not, but concedes researchers are struggling to produce new surprises and are mostly gaining incremental improvements; he adds that existing deep learning techniques alone could keep companies busy solving problems for another ten years, and points to the conference's AI Evolution Stage as the place to discuss what might come after deep learning.
A new generation of AI companies and open questions
Westerheide cites about 21,000 AI companies founded in the last five years, estimating that once purely marketing-driven ones are excluded, roughly 10,000 remain; he names the Chinese AI company Squirrel AI, which he says raised about 130 million US dollars, and Starmind among companies invited to the conference. He lists open questions the conference will address: how to get talent and data, whether to build AI in-house or acquire startups, how Europe should respond to US and Chinese competition, and questions of bias and ethics in AI decision-making. He mentions an ongoing disagreement with his wife over whether AI should have emotions, arguing that AI used in healthcare and hospitality may need to sense how people feel. He closes by announcing the next Rise of AI conference for May 13 and 14, and by outlining the day's stages: a political Future Stage on the US, China and Europe, a later panel with members of the German AI Commission, a visionary stage with science-fiction authors and transhumanists, the hands-on Applied AI Stage, the research-focused AI Evolution Stage, and the Unicorn Stage for growth-stage companies.
Quotes
This industry will be the biggest change we will see in the next decade.
Fabian Westerheide
AI is not the answer to all your problems.
Fabian Westerheide
We don't have a political agenda for this.
Fabian Westerheide
Questions answered in this talk
How has the Rise of AI conference grown since it started?
According to Fabian Westerheide at Rise of AI 2019, the event began five years earlier as a roughly 50-person meetup called The Singularity Is Near in a Berlin hackerspace and grew year by year to about 800 attendees in 2019.
How does Europe compare to the US and China in AI?
According to Fabian Westerheide at Rise of AI 2019, the US leads through its large tech companies, universities and capital, and China ranks second with a focused national strategy and roughly 130 billion US dollars invested, while Europe lags behind with no political agenda, no top-50 digital company and too few startups and capital, despite strong research institutions.
Is deep learning over?
According to Fabian Westerheide at Rise of AI 2019, deep learning is not over, but it has stopped producing dramatic new surprises, and he argues existing deep learning techniques could still keep companies busy solving real problems for about another ten years.