Welcome to Digitally Literate, issue #368. The air quality mess is starting to move down here to the southeast of the US. Thinking about those of you that have already experienced this up in the north.
Apple had its Developer’s Conference this week, and one of the key announcements was the release of Apple Vision Pro.
Many have commented about the price tag of the product. If you look at the prices of other first-gen VR equipment, I think this price makes sense, if you add the usual Apple tax.
As I watched the reviews, I was struck by how dystopian this seems. Once again, I’ve had that feeling that “this will never sell.” I had that same feeling previously with the Apple Watch…and I was dead wrong. I’m looking forward to the first time I see someone in public wearing this device.
I’ve had a lot of research and publications that have not aged well. This publication is very interesting as it provides a snapshot of AI right before we entered this frenetic pace of change.
The 2022 Expert Survey on Progress in AI (2022 ESPAI) is a survey of machine learning researchers that AI Impacts ran from June-August 2022. This is a rerun of the 2016 Expert Survey on Progress in AI that researchers at AI Impacts previously collaborated on with others. Almost all of the questions were identical, and both surveyed authors who recently published at major machine learning conferences.
Results suggest the aggregate forecast time to a 50% chance of HLMI was 37 years, i.e. 2059. High–level machine intelligence (HLMI) is when unaided machines can accomplish every task better and more cheaply than human workers. HLMI ignores aspects of tasks for which being a human is intrinsically advantageous, e.g. being accepted as a jury member. Think feasibility, not adoption.
Respondents also believe society should prioritize AI safety research “more” than it is currently prioritized. Respondents chose “much less,” “less,” “about the same,” “more,” and “much more.” 69% of respondents chose “more” or “much more,” up from 49% in 2016.
Why this matters. It’s interesting to read up on AI, ML, and related fields to get a better sense of how we got here. This survey is an indication to me that many were also caught up in the explosion of these technologies in society. There was also a fair amount of concern and responsibility as they considered the possible impacts.
Multilingual language models are new AI tools that can analyze text across languages by learning from multiple languages at the same time. These models can enable new technologies and increase access to online services in multiple languages, but they also face challenges such as errors, biases, and context sensitivity.
In a related story, Multilingual language models are being used by social media companies to detect and moderate harmful content in over 100 languages even as these models have shortcomings due to poor quality data training and disparities in the amount of data they train on in each language.
Why this matters. Governments, companies, researchers, and civil society should not overestimate these models and should invest in more transparency and accountability. Researchers from different language communities should be supported and involved in developing these models.
While the internet has provided access to a vast array of knowledge, it has also led to cognitive fracture and distraction. Similarly, while AI can be used to increase productivity, it may also distract and disrupt our ability to focus and process information. Additionally, as we delegate more cognitive and creative responsibilities to AI, we may lose the learning opportunities that come from the deep processing of information. Ezra Klein argues that the danger of over-automating cognitive and creative processes is real and worthy of serious consideration.
Why this matters. What are the potential advantages and disadvantages of using AI to produce summaries and drafts? How can AI be used to deepen human intelligence and help us be more productive?
Instagram is being used by a vast network of accounts openly promoting and purchasing underage-sex content, according to investigations by The Wall Street Journal and researchers at Stanford University and the University of Massachusetts Amherst.
This network has been connected through the platform’s systems for fostering communities. While the company claims to be improving internal controls, its role in facilitating this network raises concerns about child safety on social media.
Why this matters. As we consider the digital tools we use, we need to think about the power, purpose, and profit of the groups offering these services.
As I was digging into all of the news and reviews of Apple Vision Pro, one nugget that stuck out to me was the use of the term spatial computing. Spatial computing was defined in 2003 by Simon Greenwold, as “human interaction with a machine in which the machine retains and manipulates referents to real objects and spaces.”
At no point in the Apple announcement did they use the terms virtual reality (VR), augmented reality (AR), or mixed reality (XR).
Why this matters. What role will the user and content play in the success of spatial computing?
As the summer respite begins, I’m taking time to identify some things I’d like to do but gave up as life and adulting get in the way. For me, part of this involves taking some time to play video games again. This time, I can do it with my kids. 🙂
One lesson of the digital age is that more is not always better. More emails and more reports and more Slacks and more tweets and more videos and more news articles and more slide decks and more Zoom calls have not led, it seems, to more great ideas.
Cover Photo CC BY using DreamStudio.AI