HUMAN + MACHINE = SUPERPOWERS — notes from DLD18
Visiting a new conference is always slightly unnerving at first — frankly, will this be useful or will it be waste of my time. But DLD 18 Munich has hit all the right notes in the first morning. My energy is up, my notebook is filling up and I’m looking forward to more.
Day one started with volley couple of fascinating sessions, the public policy head for Facebook standing his ground in the face of political and publishing industry ire, Andrew Keen talking with one of Germany’s biggest media companies about the future of the internet, and a panel kicking around whether apps or agents will be more important in the coming years.
Then came Paul Daugherty, CTO of Accenture to talk AI. He is one of those fast-talking guys who wants to tell you everything and who you almost wish could talk even faster and download it all in 25 minutes on stage. As it, what he covered in terms of insights and questions was provocative and compelling. I’ll just list them out here for now:
- AI is the fastest growing business:He’s never seen anything grow so fast.
- Job skills are more of a problem than job destruction: Something like 6 million people are unemployed in the US and there about the same number of jobs vacant.
- We need to train many more people in AI: Estimates that there are only 10,000 people worldwide with the right level of skills and knowledge in AI are probably correct — we’re going to need a lot more.
- Tech can help people work better: If we look at how work happens we can implement tech to help people work better. He cited a major US manufacturer that is using augmented reality (AR) headsets to help employees operate machinery better (I imagine a bit like the virtual hands teaching piano learners how to hit the right notes).
- Business processes are the main issue for AI: Daugherty calls the AI wave of innovation in this area — Business Process 3.0.
- New jobs are being created in training AI: AI needs to be trained and supervised — for instance to make sure that customer service bots and interfaces are reflecting the tone and values of the company in their interactions and decisions.
- Leadership is a key issue: CEOs and board level people need to learn about AI. Echoing our own digital leadership mindset work at Brilliant Noise, Daugherty also said that it was more than just AI they needed to learn about: “People need to invest in skills for leaders to move through [successive] generations of technology.” 85% of leaders say AI is something they will be investing in but only 3% are investing in training.
- Silo-busting is a must: AI is most effective when it is at the core of new business processes, not added to the periphery.
- Chief Artificial Intelligence Officers (CIAOs) are a good idea: The CIAO should own the three areas key to making AI successful in an organisation: 1. Data, 2. Talent, 3. Responsible use.
- Responsible use of data and AI: There needs to be oversight in this area to watch out for the unintended consequences of AI and automation. There will alway be hidden biases in algorithms.
- “Data is fuel for AI”: Companies that will find it difficult to succeed with AI are those with unstructured and disconnected data sets. AI needs data to be effective.
- Absorbative capacity: A useful term from economics, absorbative capacity refers to how quickly an economy can take advantage of innovation.
- Productivity needs better measurement: If we can measure it better we can figure out where to boost it.
Paul Daugherty has a book — Human + Machine — on this topic coming out in March — it’s going to be top of my reading list as soon as I can get a copy.