Last Week in AI #90

Impacts of AI systems' decisions, future of deep learning, and more!

Nov 9, 2020

Last Week in AI #90 — Image credit: Meredith Whittaker / Boston Globe

Mini Briefs

Who am I to decide when algorithms should make important decisions?

As more AI systems are being used in production systems and making decisions that impact our daily lives, it’s important to start incorporating the voices of the people that these decisions directly impact. The grocery platform Shipt recently implemented a new algorithm that automatically determines workers’ schedules and wages. Many workers complained about reduced wages as well as unpredictable scheduling. The decision about whether or not AI-powered decision systems are used is solely determined by corporations or government institutions, where the primary benefactors are themselves. Furthermore, claims of “fairness” and transparency are often unsubstantiated and presented without evidence or verification. The article explores the voices and perspectives of people directly impacted by AI systems and advocates for the inclusion of these people into discussions about the appropriate use of AI.

AI pioneer Geoff Hinton: “Deep learning is going to be able to do everything”

Geoffrey Hinton, a Turing Award Winner and an AI Pioneer, believes that deep learning will be sufficient to achieve AGI (artificial general intelligence). In his interview with MIT Technology Review, Hinton states that while key conceptual breakthroughs are necessary (e.g., new architectures similar to transformers in NLP), bridging these conceptual gaps and scaling up existing architectures and data should be sufficient to achieve the goals of AI research.

It’s worth noting that this opinion might not be shared by everyone working in AI. However, having worked on neural networks before they became popular, Hinton comments on his track record of contrarian views:

Well, my problem is I have these contrarian views and then five years later, they’re mainstream. Most of my contrarian views from the 1980s are now kind of broadly accepted. It’s quite hard now to find people who disagree with them.

Podcast

Check out our weekly podcast covering these stories! Website | RSS | iTunes | Spotify | YouTube

News

Advances & Business

Ocado adds to robotics capability with $287 million acquisitions - “British online supermarket and technology group Ocado said it would buy two robotics companies for a total of $287 million and upgraded its full year earnings outlook on the back of strong trading at its retail joint venture with Marks & Spencer.”
Walmart is reportedly giving up on shelf-scanning robots in favor of humans - “Retail robots that can scan shelves and update inventory have been one of the most visible faces of automation in recent years, but the success of such machines is far from guaranteed.”
Where to patrol next: ‘Netflix’ of ranger AI serves up poaching predictions - “The PAWS AI system, developed out of Harvard University, uses data about past poaching and game theory to predict where rangers are most likely to find poaching activity next.”

Concerns & Hype

Canada crawling toward AI regulatory regime, but experts say reform is urgent - “On Thursday, privacy watchdogs revealed that five million images of shoppers’ faces were collected without their consent at a dozen of Canada’s most popular malls.”

Analysis & Policy

The near-future of AI and automation is a hybrid workforce, not mass job loss - “AI and automation are receiving a boost during the coronavirus pandemic that in the short term is creating a new hybrid workforce rather than destroying jobs outright.”

Expert Opinions & Discussion within the field

How many AGIs can dance on the head of a pin? - “In the absence of a good and unequivocal definition of intelligence, how can we discuss AGI and superintelligence?”
What AI Can Learn From Parents - “Like raising children, training an artificial intelligence means balancing obedience and independence.”
Teaching AI agents to communicate and act in fantasy worlds - “In recent years, artificial intelligence (AI) tools, including natural language processing (NLP) techniques, have become increasingly sophisticated, achieving exceptional results in a variety of tasks.”
How I got 10k post karma on reddit with (and without) fast.ai - A reddit user details how to build a language bot to write popular articles on reddit.

That’s all for this week! If you are not subscribed and liked this, feel free to subscribe below!

Last Week in AI