At Rise of AI 2016 in Berlin, host Fabian Westerheide opens the event's second edition with a status update on artificial intelligence and its effect on humans. He notes the event grew from 20 attendees to about 150 in two editions, and announces he has just become the coordinator for AI at the German Startup Association, inviting AI companies to reach out for policy-level representation. He argues machine learning and AI have existed since the 1950s-70s but improved sharply through more processing power, deeper neural networks, better natural language and image recognition, and the rise of big data and the internet. He lists tasks where narrow AI already matches or beats humans, from games, drones and cars to writing emails, trading stocks, detecting fraud, and instant medical diagnosis. Drawing on roughly 250 audience answers to what makes humans unique, he argues traits like creativity or planning are already replicated by machines, while consciousness, emotions and ethics remain unsolved. He closes with a "wealth index" showing rising lifespan and wealth and falling working hours since industrialization, concluding AI should not be feared.
Key takeaways
- Westerheide had just become AI coordinator for the German Startup Association and invited AI companies to reach out.
- Neural networks grew from roughly one layer in the 1960s-70s to about a thousand layers by 2016.
- He lists instant medical diagnosis and supply-chain optimization as emerging strengths of narrow AI in 2016.
- An audience survey of 250 answers found consciousness, emotions and ethics as the traits AI had not replicated.
- Global wealth rose about 600 percent and lifespan about 150 percent since the Industrial Revolution, per his wealth index.
Detailed notes
Opening and a new AI policy role
Westerheide welcomes the audience to the second Rise of AI, noting growth from 20 attendees at the first edition to about 150 at this one, with an expectation of 300 next time. He introduces himself as running a venture capital firm focused on AI in Berlin and organizing Rise of AI twice a year together with his wife Veronica from Pandora Events. He announces that he has just become the coordinator for artificial intelligence at the German Startup Association, inviting AI companies and anyone interested in discussing AI at a policy and economic level to contact him, describing the role as working toward the path to singularity.
The state of AI in 2016
Westerheide frames the "gaffa economy" of Google, Amazon, Apple and Facebook as driving progress, and notes machine learning has existed since the 1950s-70s through repeated AI winters, but neural networks that once had one layer now have a thousand, enabled by more processing power. He points to progress in natural language processing and visual and audio recognition, citing Google's cat-picture recognition as an example, and to the rise of big data as corporations collected data and needed machine learning to use it. He credits the internet as a precondition for the current wave, noting rising AI startup funding and corporations acquiring AI teams.
What narrow AI already does better than humans
Westerheide describes today's AI as narrow but already superior to humans at specific trained tasks. Examples include beating humans at Go, chess and card games, flying drones and driving cars, organizing meetings and booking travel, writing songs and drawing pictures, trading most stock market volume, detecting financial fraud, writing personalized sales emails and handling customer support, and diagnosing illness faster than many doctors. He also describes AI optimizing complex production and supply-chain processes that previously required decades of human experience, and notes machines work continuously without striking or needing breaks.
What makes humans unique
Westerheide shares results from asking the audience what makes humans unique, having collected over 250 answers. He groups traits like creativity, taste, critical thinking, visionary thinking and long-term planning as already replicated by machines in non-human ways, citing Netflix's content recommendation and production decisions as an example. What remains unsolved, he says, are conscience, compassion, emotions, greed, love, hate, ethics, morality, humor, curiosity and insecurity, none of which he had seen any startup attempt to solve. He concludes that feeling, not work, is what defines human uniqueness.
The human wealth index
Westerheide presents a "human wealth index" showing little change for 2.5 million years of hunter-gatherer life until the Industrial Revolution, when machine use accelerated human wellbeing. He cites a 150 percent increase in lifespan over the past 150 years, a 600 percent increase in global wealth, and a drop in average working hours from 75 hours per week in 19th-century France to roughly 35-40 today. He concludes that embracing technology produced these gains and that AI should be approached the same way, to live longer, be more intelligent, and have more free time.
Quotes
We don't have to fear AI, we take AI to live longer, to be more intelligent, to be more happy, to have more free time.
Fabian Westerheide
AI is better than we humans are in a very specific task, once they are trained.
Fabian Westerheide
Feeling, I think, is being unique. Having emotions is unique.
Fabian Westerheide
Questions answered in this talk
What tasks could AI already do better than humans in 2016?
According to Fabian Westerheide at Rise of AI 2016, AI already beat humans at games like Go and chess, flew drones, drove cars, traded most stock market volume, detected fraud, wrote emails, and diagnosed illness faster than many doctors.
What did Rise of AI attendees say makes humans unique compared to machines?
Fabian Westerheide reports that out of over 250 audience answers, traits like creativity and planning were considered already replicated by machines, while consciousness, emotions, ethics and morality remained unsolved.
How has technology changed human wealth and working hours historically?
Fabian Westerheide cites a roughly 150 percent increase in human lifespan and 600 percent increase in global wealth since the Industrial Revolution, alongside a drop in average working hours from about 75 per week in 19th-century France to roughly 35-40 today.