Search
There is a significant gap in research about Canadian data collection activities on a granular scale. This lack of knowledge regarding data collection practices within Canada hinders the ability of policymakers, civil society organizations, and the private sector to respond appropriately to the challenges and harness unrealized benefits.
So true. This looks like an interesting series from the great team at Brookfield.
In the next three years, as many as 120 million workers in the world’s 12 largest economies may need to be retrained or reskilled as a result of Artificial Intelligence (AI) and intelligent automation.
cf. Lee Se-Dol.
This is according to the latest IBM Institute for Business Value (IBV) study, titled The Enterprise Guide to Closing the Skills Gap.
Seems like an interesting guide. This metric surprised me:
In 2014, it took three days on average to close a capability gap through training in the enterprise. In 2018, it took 36 days.
I didn’t know this measure existed, but I can see the utility. As knowledge work grows ever more specialized, this time-to-capability can only grow.
The over- and misuse of AI is one of my biggest tech pet peeves. It truly is evil to tack the AI term onto the description of most products. It also damages the long-term potential of AI by corrupting what it means—especially for the everyday people who aren’t involved or invested in building these tools, but who will use them (or be used by them).
Arvind Narayanan on Twitter:
Much of what’s being sold as “AI” today is snake oil. It does not and cannot work. In a talk at MIT yesterday, I described why this happening, how we can recognize flawed AI claims, and push back. Here are my annotated slides: https://www.cs.princeton.edu/~arvindn/talks/MIT-STS-AI-snakeoil.pdf
Key point #1: AI is an umbrella term for a set of loosely related technologies. Some of those technologies have made genuine, remarkable, and widely-publicized progress recently. But companies exploit public confusion by slapping the “AI” label on whatever they’re selling.
Key point #2: Many dubious applications of AI involve predicting social outcomes: who will succeed at a job, which kids will drop out, etc. We can’t predict the future — that should be common sense. But we seem to have decided to suspend common sense when “AI” is involved.
Key point #3: transparent, manual scoring rules for risk prediction can be a good thing! Traffic violators get points on their licenses and those who accumulate too many points are deemed too risky to drive. In contrast, using “AI” to suspend people’s licenses would be dystopian.
“In the five years that weʼve had to asses the effect [the Gigafactory has] had on the workforce, on the community, I think there have been these ramifications that we talk about in the episode that nobody was really prepared for,” Damon said in an interview with The Verge. “Like, we knew there was going to be an issue with housing, which other cities are experiencing, too. But thatʼs become super critical.”
Side-effects of growth are not a new problem, but the massive initiatives we’re seeing recently might spark new varieties of old issues.
Land that became too toxic for people to farm and live on after the 2011 meltdown at the Fukushima Dai-ichi Nuclear Power Station will soon be dotted with windmills and solar panels.
The Fukushima disaster unfolded as an incredible story of systemic response to new scales of tragedy. Take, for instance, the Skilled Veterans Corps: a group of elderly volunteers who helped with cleanup, knowing that the damaging radiation would have less impact on their lives than it would on younger volunteers.
Now Fukushima’s next chapter is evolving as an example of systemic creative destruction, as new opportunities are unlocked by the collapse of the region’s previous energy strategy.
The South Korean Go champion Lee Se-dol has retired from professional play, telling Yonhap news agency that his decision was motivated by the ascendancy of AI. “With the debut of AI in Go games, Iʼve realized that Iʼm not at the top even if I become the number one through frantic efforts,” Lee told Yonhap. “Even if I become the number one, there is an entity that cannot be defeated.”
Wow. Perhaps the first real example of “AI took my job?”
In this research article, the authors point out that the cycles of translation from English to the language of the context and back again can be costly and inconvenient. But, they point out three benefits to investing in translation and multi-lingual research spaces.
First, the authors argue that disseminating the results of research in local languages not only makes your research accessible to stakeholders, but it also helps stakeholders value all research more. They write:
Translations are expensive and time-consuming, so a large part of our work stays in English unavailable to the local stakeholders, who may have participated in the research process. This is an issue not only because it reduces their possibilities to learn from the systematized outcomes of the processes in which they participate, but because it reduces their perception of the value of research. When stakeholders feel that researchers write exclusively for other foreign researchers, their readiness to support and fund research may decrease.
The second benefit:
Second, academics who don’t read English may find it difficult to continue building on knowledge published only in that language.
This takeaway is obvious. So many publications are translated to English, but the reverse is rare.
Third, and by no means least, naming complex issues or ideas only in English impoverishes other languages. When we forsake finding a word for a particular concept or idea in a given language, we impoverish that language.
This is quite insightful. Language is intrinsic to organizational learning. If the concepts advanced in our research are never introduced to the local language, then it may be impossible for that learning to take root.
The authors recognize a fascinating tension in this work. They demonstrated the possibility of multi-language research spaces via a virtual research commons for their project.
What we have learned from working with different languages and acknowledging them during the full research cycle, including the dissemination stage, is that they are time- consuming, costly and even a bit messy and uncomfortable. For example, in the case of the virtual space above, some participants complained that having to find their own language among texts written in other languages begs an extra effort from them and slows them down. However, the alternative is renouncing inclusion and plurality, which is at odds with the challenge faced by academia to address complex societal problems.
There is a cost to complexity, but solution spaces need to be more complex than the problems they’re resolving.1
See Ashby’s Law of Requisite Variety; http://pespmc1.vub.ac.be/ASHBBOOK.html, p. 207 ↩︎
Every part of this trial sounds made up. They should just air it in lieu of a Good Fight episode. Elizabeth Lopatto’s writeup is worth worshipping.
Spiro then coined the worst acronym I’ve heard in years, and I edit stories about aerospace so I know from bad acronyms. It is: JDART, for joking, deleted, apologized-for, responsive tweets.
Incredible.
But there’s at least one abstract takeaway that’s interesting to me:
At this point, Wood tried to enter an email exchange into evidence, resulting in a great deal of confusion on Judge Wilson’s part about how email reply chains work. (You read from the bottom.)
[…]
At this point, the “pedo guy” Twitter thread was entered into evidence, and the befuddled court had to be told that the reply chains work the other way on Twitter — the first tweet is at the top, and the last tweet is at the bottom.
Yet another example of the ways in which the world’s accelerating faster than many institutions can keep up.
Incredible achievement, but it makes me wonder—what are the .2% of humans doing differently?
These stories of AI achievement are sure to proliferate in the coming years. By focusing on those people who are still able to think around machine learning strategies, we might learn something about how humans and machines can best complement each other.
The halo effect is essentially how positive—but irrelevant—traits influence our perception of what the thing with the halo actually says or does. These authors explored how charities manifest the halo effect on their websites, and find evidence for four varieties of halo effect.
this study employs charity websites as a multi-attribute donation channel consisting of three attributes of information content quality (mission information, financial information, and donation information) and four attributes of system quality (navigability, download speed, visual aesthetics, and security). Based on the proposed framework, this study proposes four types of halos that are relevant to charity website evaluation —collective halo (attribute-to-attribute), aesthetics halo (attribute-to- dimension), reciprocal-quality halo (dimension-to-dimension), and quality halo (dimension-to-dimension)
