I really enjoyed the @nberpubs Economics of AI conference in Toronto. Very nice mix of papers and excellent discussants.
Kudos to @professor_ajay, @joshgans, @avicgoldfarb and @ce_tucker for organising.
Some highlights below. Chad Jones formalised the dilemma between AI-powered growth and existential risks. An interesting result is that AI systems that extend life expectancy increase the attractiveness of investing in AI-intensive growth despite the risks it could create.
economics.princeton.edu/wp-content/upl…

"Modelling an evolving economy"
Last Friday, @ESCoEorg & @DigiCatapult brought together researchers and policymakers to discuss how we can use new sources of data and analytical methods to understand & steer a fast-changing economy.
Here is a thread with some highlights. 1. In the first keynote, @MDelgadoSevilla presented her work mapping clusters, the supply chain economy and inventor inclusion.
One of her most interesting findings is that innovative firms benefit from clustering *specially* when there is a negative shock.

"Superhuman science"
This paper explores the potential impact of AI on scientific fields where innovation is about searching big spaces of possibility (e.g. new drugs, materials, architectures to build better AI systems...).
brookings.edu/wp-content/upl…
#econai ht @mattsclancy 1. AI systems trained on historical data can help predict which opportunities have more potential, improving prioritisation and expected returns from innovation.
But downstream blockers (e.g. around testing of the ideas that are developed) could constrain these benefits.

Today we are hosting @smpfotenhauer & @Jackstilgoe to discuss their politics of scaling paper & its implications for @nesta_uk. Here is an older thread with some reflections about it:

https://twitter.com/jmateosgarcia/status/1450362833313714177

Some implications:
1. Be inclusive when exploring alternatives
2. Avoid a bias towards scalable solutions
3. Secure consent for experiments & ensure that they are safe & beneficial in the 'test site'
...

"Any site where scaling is made to look easy should raise red flags about a likely lack of comprehension or inclusiveness of perspectives"
The Politics of Scaling paper highlights the risks of scaling efforts to tackle big societal challenges.
journals.sagepub.com/doi/full/10.11… 1. Three shared features of scaling efforts give rise to potential disfunction:
a. Solutionism: societal challenges are fixable
b. Experiments: We can deploy local solutions globally
c. Future-oriented value: scaling focuses on future benefits over present ones.

Some observations about the UK AI strategy:
1. I like that it presents public funding for R&D as a mechanism to help steer AI in a societally beneficial direction. This sets it in contrast with other national strategies that put R&D and public value / ethics in silos. 2. I like its emphasis in the importance of workforce diversity for building better and more inclusive AI systems. However, it neglects other (correlated) types of diversity e.g. disciplinary, technological, institutional that are also important (cf. arxiv.org/abs/2009.10385)

Jer Thorp’s Living in Data starts strong: “our projects became less about finding answers in data and more about finding agency, less about exploration and more about empowerment.”

Looking forward to read where it goes with this idea. “The lesson I learned was to treat data and the system it lived in not as an abstraction but as a real thing with particular properties, and to work to understand those unique conditions as deeply as I can” ~ use data not to scale up away from the thing, but back into it.

Mission Measurable: We just published a working paper and blog with some prototype indicators for innovation missions.

Brief thread:

nesta.org.uk/blog/mission-m…

papers.ssrn.com/sol3/papers.cf…

Perhaps FYI @jameswilsdon @IIPP_UCL

Acknowledgement: this is a @EuritoH2020 output. 1. We argue that innovation missions need new indicators able to capture emergence, diffusion, crossover and diversity. Otherwise, how will we know if they work?

Mission oriented innovation policies are after all not without challenges & risks (eg see st-andrews.ac.uk/business/rbf/w…)

[Highlights of day 2 of the Economics of AI @nberpubs workshop] Sonny Tambe presented a cool paper using LinkedIn data to recover firm investments in intangible AI-related assets, finding it concentrated in 'superstar' firms (FAO @stianwestlake)
papers.ssrn.com/sol3/papers.cf… 2. @danielrock presented his work about the value of engineering / AI. I discussed it here:

https://twitter.com/JMateosGarcia/status/1176369435805913089

Fantastic keynote about policies to generate radical innovation to kick off #IGL2019. How do you design institutions to be able to do this? Some notes:
1. Control your failure (error) rate *upwards*. A perfectly successful portfolio isn’t supporting radical innovation. 2. Orchestrate technological ecosystems with deliberation. This is what DARPA project managers do, arranging technologies like chess pieces (see Erica Fuchs’ slide)

Today we launch Innovation Mapping Now, a report about our @nesta_uk work using new data sources, analytics and visualisations to inform innovation policy.
Summary thread ensues [1/7]
nesta.org.uk/report/innovat… We start with the policy problem: a disconnect between technological innovation and growth/sustainability/wellbeing. Traditional innovation policy instruments don't seem enough to address it. A new wave of activist, directional & holistic policies emerge as an alternative [2/7]

Some interesting from an intense two-day innovation metrics review workshop in Canberra:
1. Everybody wants to measure culture and incremental innovations inside businesses: development, extension & continuous improvement.
[1/6] 2. Qualitative case stories have a place in quantitative monitoring framework. I like this 'telling stories with numbers' NIH example by M Ann Feldman [2/6]