1 – A Review of Nick Bostrom’s Superintelligence: Paths, Dangers, Strategies Leaves More Questions
A book review of Nick Bostrom’s Superintelligence: Paths, Dangers, Strategies, looks at Bostrom’s warnings and potential problem-solving strategies for the rise of the so-called superintelligence (i.e. a self-improving AI) against a historical backdrop of scientific accomplishments and blunders. Bostrom warns of the risks posed by the onset of intelligent machines, a view that has recently been reinforced by tech icons like Elon Musk, Bill Gates, Stephen Hawking, and others. While the idea of a “superintelligence” has many roots in Silicon Valley and prominent thinkers, including Google’s Director of Engineering Ray Kurzweil, Bostrom’s book has successfully brought earlier dismissed concepts into the mainstream intellectual discourse. Much of that conversation centers on the “control problem” of how to prevent machines from determining the fate of mankind, including current AI-enhanced technologies that pose dangerous potentials for existing 21st-century threats.
(Full article on Bulletin of the Atomic Scientists)
2 – Scientists and Tech Gurus Write Open Letter Warning Against Autonomous Weapons Arms Race
Leading world scientists and tech experts banded together in an open letter that warns of a “virtually inevitable” global arms race with weapons using artificial intelligence. Unlike a remotely-piloted drone, autonomous weapons would have complete control over target selection and killing. In an interview with NPR, AI Researcher Stuart Russell discerns between two arguments paralleling the development of these machines. There’s the engineering argument – are we capable of building a machine that is better able to make such life or death decisions than humans; then, there’s the overwhelming argument against the production of millions of devices that, unlike nuclear power, are available at low costs to almost any motivated person or group of people. The United Nations is currently working on a treaty that would ban autonomous weapons.
(Full article and interview on NPR)
3 – Artificial Intelligence Makes Moves in Genetics-Based Diagnostics
AI is making headlines in the health industry, and one way its being tagged is for more precise diagnoses based on individual genetics. Hypothetically, AI could be used to diagnose an individual’s entire genome and environmental history, diagnose a disease, and suggest treatment. Dr. Niven Narain, CTO for Biopharma stated,
“I think AI is what is going to drive this voluminous amount of information into going from data to knowledge, and from knowledge to products.”
Through the existing AI system EMERGENT, a group of researchers recently discovered five new biomarkers that could be used to develop a treatment for glaucoma. The system identified this information using information from 2,300 individuals, sourcing information on more than 600,000 specific DNA sequences and knowledge of specific genetic interactions. Researchers also note the potential for pharmaceutical companies to use AI to decrease time and costs in developing a single drug, which typically takes 14 years and costs $2.6 billion. This type of functional technology is still decades away in the making.
(Full article and interview on International Business Times)
4 – Cubic is the New Digital Assistant Taking Business and Personal to New Levels
Cubic Robotics, a technology company based in Palo Alto, California, is on the verge of introducing an artificially intelligent personal assistant that will be able to connect to all of a person’s technology devices, apps and services. The technology behind Cubic consists of two hardware platforms – the “home cube” and the “power badge” – allowing for mobility. Cubic will be able to recall lifestyle patterns and operate from afar (turning off forgotten lights, for example). The technology is also meant to be interactive and adaptively intelligent, familiarizing itself with your habits, interests and persona, and updating its cloud-based “mind” accordingly. Look for the first artificially intelligent digital assistant to be released to the public in January of February of 2016. Investors on IndieGoGo are already purchasing the first batch of Cubics to be released.
(Full article and interview on Huffington Post)
5 – Startup Nnaisense on a Mission to Build Neural Network-Based Artificial Intelligences
The new startup Nnaisense, founded by Jürgen Schmidhuber, is on a mission to to “build large-scale neural network solutions for superhuman perception and intelligent automation, with the ultimate goal of marketing general-purpose neural network-based Artificial Intelligences.” Schmidhuber is a pioneer in the area of recurrent neural nets (RNNs), technology that functions as a form of short-term memory. RNNs is part of the AI trend of “deep learning”. An RNN has the capability to examine one thing, while keeping another idea in mind, and can then use this information to help analyze the other.
(Full article on Wired)