Connect with us

Artificial Intelligence

Research Priorities for Artificial Intelligence – Open Letter

Published

on

The Future of Life Institute, a research and outreach organization working to mitigate existential risks facing humanity, currently focusing on potential risks from the development of human-level artificial intelligence, has issued an open letter to the Artificial Intelligence (AI) community.

// -- Discuss and ask questions in our community on Workplace.

There is now a broad consensus that AI research is progressing steadily, and that its impact on society is likely to increase. The potential benefits are huge, since everything that civilization has to offer is a product of human intelligence; we cannot predict what we might achieve when this intelligence is magnified by the tools AI may provide, but the eradication of disease and poverty are not unfathomable. Because of the great potential of AI, it is important to research how to reap its benefits while avoiding potential pitfalls. [W]e recommend expanded research aimed at ensuring that increasingly capable AI systems are robust and beneficial: our AI systems must do what we want them to do. [W]e believe that research on how to make AI systems robust and beneficial is both important and timely, and that there are concrete research directions that can be pursued today.

An attached research priorities document gives many examples of such research directions that can help maximize the societal benefit of AI. This research is by necessity interdisciplinary, because it involves both society and AI. It ranges from economics, law and philosophy to computer security, formal methods and, of course, various branches of AI itself.

Fear of Smarter-than-Human AI

Robot womanThe open letter has been signed by a who-s-who list of experts in advanced AI. Notably, the list of signatories include renowned cosmologist Stephen Hawking, Tesla Motors and SpaceX founder Elon Musk, and Oxford philosopher Nick Bostrom, author of the recently published book “Superintelligence: Paths, Dangers, Strategies” (Here is my full review of the book). After the publication of Bostrom’s book, Hawking and Musk expressed fears about the future of AI. In August, Elon Musk tweeted that artificial intelligence could be more dangerous than nuclear weapons and in October, likened it to “summoning a demon.” Stephen Hawking told the BBC in December that AI could “spell the end of the human race.”

// -- Become a yearly Platinum Member and save 69 USD and get access to our secret group on Workplace. Click here to change your current membership -- //

Also read: Artificial Intelligence – A Military Roadmap

Bostrom defines superintelligence as something far smarter than us, not in the provincial sense that Einstein is smarter than the village idiot, but in the real sense that Einstein (or the village idiot – the difference is utterly irrelevant on this scale) is smarter than a beetle. Bostrom thinks that the first human-equivalent AI able to learn and improve itself could be developed sometime in this century, with the possibility of a very fast transition to superintelligence soon thereafter. A superintelligent and hostile AI could, as Musk and Hawking fear, eliminate humanity.

The signature of Musk and Hawking may be interpreted as expressing confidence that adopting the recommendations in the open letter would diminish the risks of uncontrolled development of AI. You can sign the letter here of you want. I didn’t sign it (yet), because I think that important progress in AI, including the development of smarter-than-human AI and superintelligence, can only emerge from free, spontaneous and unconstrained research. I don’t disagree with the open letter or the research priorities document, but setting common priorities is not the aspect of AI research that I find more interesting at this moment – I prefer to let a thousand flowers bloom.

The mission of he Future of Life Institute is to catalyze and support research and initiatives for safeguarding life and developing optimistic visions of the future, including positive ways for humanity to steer its own course considering new technologies and challenges. The Institute received seed funding from Skype co-founder Jaan Tallinn and Matt Wage.

Images from Shutterstock.

Important: Never invest money you can't afford to lose. Always do your own research and due diligence before placing a trade. Read our Terms & Conditions here.



Feedback or Requests?

Artificial Intelligence

Bitcoin Giant Bitmain Enters the High Stakes AI Race

Published

on

Artificial Intelligence

The Sophon, named for a fictional proton-sized supercomputer, could be the tool to train neural networks in data centers worldwide. It is the latest project being developed by Bitmain Technologies Ltd., the bitcoin mining giant that has carved out a dominant position in bitcoin mining.

// -- Discuss and ask questions in our community on Workplace.

Such chips, called application-specific integrated circuits (ASICs), could unleash a new wave of distributed computing, according to Michael Bedford Taylor, a University of Washington professor who studies bitcoin mining and chips.

Sophon is due to debut before the end of the year.

Bitmain Has The Know-How

Bitmain has the background to play a role in the expanding artificial intelligence industry. The company designs the silicon that goes in bitcoin mining equipment, assembles the machines and sells them worldwide, in addition to its own bitcoin mining operation and the ones that it manages for other mining pools.

// -- Become a yearly Platinum Member and save 69 USD and get access to our secret group on Workplace. Click here to change your current membership -- //

Bitmain’s founders are not averse to playing a spoiler role.

Jihan Wu, the co-founder of Bitmain, supports the New York Agreement that seeks to double the bitcoin block size under the SegWit2X proposal, a move that some in the bitcoin community view as an attempt to give the miners control over bitcoin.

Some also believe Wu was behind the recent bitcoin split known as bitcoin cash, which at least one of Bitmain’s miners supported, a contention that Wu has denied. Wu points out that he was among the supporters of Bitcoin Unlimited, an earlier bitcoin scaling proposal that did not get activated.

Why Wu Supports Forks

Wu nonetheless said splits should be allowed. He said a fork is inevitable since people in the bitcoin community do not agree on how to best scale bitcoin.

Wu met Micree Zhan, Bitcoin’s co-founder, when Zhan was running DivaIP in 2010, a company that made a device that allowed a user to stream a TV show on a computer screen.

In 2011, Wu needed a chip designer to build a mining operation and approached Zhan. Zhan first designed an ASIC to run SHA-256, the cryptographic calculation used in bitcoin, at maximum efficiency. It took him six months to finish the job. His first rig, Antminer S1, was ready in November 2013.

Bitmain felt the sting of the 2014 Mt. Gox meltdown. But by 2015, bitcoin’s price bottomed out and later recovered. In the meantime, Bitmain introduced its Antminer S5.

Bitmain now employs 600 people in Beijing.

Also read: Bitmain clarifies its ‘bitcoin cash’ fork position

Ready To Take On Google

Bitmain has since developed a deep learning chip with improved efficiency. Users will be able to build their own models on the ASICs, enabling neural networks to deliver results at a faster pace. Google’s DeepMind unit used this technique to train its AlphaGo artificial intelligence.

Bitmain plans to sell the chips to any company looking to train its own neural nets, including firms like Alibaba, Tencent and Baidu. Bitmain could build its own data centers with thousands of deep learning rigs, renting out the computation power to clients the way it does with bitcoin mines.

Professor Taylor said companies like Bitmain that have excelled in bitcoin mining could take on the Googles and Nvidias since they have developed the skills to survive in an ultra-competitive and highly commoditized industry, and have the system level design expertise and the ability to reduce data center costs.

Important: Never invest money you can't afford to lose. Always do your own research and due diligence before placing a trade. Read our Terms & Conditions here.



Feedback or Requests?

Continue Reading

Artificial Intelligence

Dutch Police Use Augmented Reality to Investigate Crime Scenes

Published

on

Dutch police are undertaking an experiment to see if augmented reality can help officers at a crime scene, according to the New Scientist.

// -- Discuss and ask questions in our community on Workplace.

Using an AI system video from the body cameras worn on the officers at the scene will relay back to experts who can guide the officers by making virtual notes which the officers will be able to see via a smartphone or head-mounted device.

Dragos Datcu, principal researcher at augmented reality (AR) company Twnkls in Rotterdam, the Netherlands said:

We now have good enough software and hardware to use augmented reality at crime scenes.

What’s great about the new AI system is the fact that experts can get involved with what the crime scene investigators are doing regardless of where they are located.

// -- Become a yearly Platinum Member and save 69 USD and get access to our secret group on Workplace. Click here to change your current membership -- //

So by viewing the footage that is sent from a camera on the police vest, a chemical specialist in one location can view it while a forensic scientist in another location can too. The system is similar to the popular Pokémon Go smartphone game that has grabbed the attention of millions of people around the world.

Not Suitable for Making an Arrest

However, while the technology may prove beneficial in providing an extra pair of eyes for investigating crime scenes, when it comes to making an actual arrest the technology is not suitable for that just yet.

According to Nick Koeman, innovation adviser from the National Police of the Netherlands, the officers undertaking the AI system trial found the extra information distracting.

Of course, some may simply say that ensuring a complete team is at the scene of a crime would be more beneficial for an investigation; however, that is not always possible due to budget cuts and time constraints.

As such the use of an AI system that can cut down on the number of people involved at a crime scene without sacrificing on the required thoroughness could potentially provide the answer that many police departments are searching for.

Not only that, but by reducing the number of people at a scene it cuts the potential possibility of contaminating evidence. The use of AI gives people the chance to assess the evidence and discover additional clues without being at the crime scene.

AI could also help in court cases by helping to recreate a scene for a jury, but as Michael Buerger, professor of criminal justice at Bowling Green State University in Ohio states, legal challenges are likely to raise when augmented reality (AR) is used in the courts.

Featured image from Shutterstock.

Important: Never invest money you can't afford to lose. Always do your own research and due diligence before placing a trade. Read our Terms & Conditions here.



Feedback or Requests?

Continue Reading

Artificial Intelligence

Apple Patent Reveals Siri-Assisted iMessage P2P Payments Platform

Published

on

Apple has filed a patent application for a “virtual assistant in a communication session.” On first glance, application 14/713,410 might scare developers of chat bots and the like, but Apple is specific in the scope of their claims – Siri, their virtual assistant platform, will be able to directly communicate with either user in an iMessage conversation and then be able to act on the instructions given. The other participant of the conversation will not see messages intended to Siri, nor messages sent from Siri to the user activating her. This is somewhat different from current implementations of chat bots and virtual assistants in chat programs, in that the classical chat bot for, say, IRC, would be accessible by most of the users of the chatroom.

// -- Discuss and ask questions in our community on Workplace.
Siri Sharing Location Information Between iMessage Participants

Source: USPTO

Siri will also be able to privately give information requested of the other party, even if the other party has not provided it. In the above image, Siri lets the user know that everyone will arrive within five minutes, clearly flexing the GPS data available to her from the other user’s iPhone. One would assume a massive update to the iMessage user agreement will be necessary, and the privacy implications of voluntarily allowing an AI to be involved in every conversation are clear. However, as shown, the default is for the user to give permission for their location information to be shared. If Siri is ever compromised, so too could every conversation on the famously secure iMessage protocol.

Siri Scheduling Meetings Between Participants

Source: USPTO

// -- Become a yearly Platinum Member and save 69 USD and get access to our secret group on Workplace. Click here to change your current membership -- //

However, Siri is not involved in conversations until she is summoned. She is added as another participant when the user summons her or, presumably, adds her manually. As seen in the image above, she can schedule meetings in addition to her other functionalities. However, perhaps most interesting, and most appetizing for malicious hackers, is the prospect shown in the below image. Users who have their financial details linked to their Apple account (which is to say, most or all users) are able to send money via iMessage at the stroke of a message.

Siri Faciliating P2P Payments in iMessage

Source: USPTO

As you can see, Siri suggests giving the funds via cash, likely due to the small amount. According to the patent application, Siri first figures out what options are available to both party, and then presents options. Both users appear to use Bank of America and Paypal, but Siri recommends cash. Both the illustrations and their descriptions in the application make it unclear which option the user selects, although the logo to the left of the success message indicates that cash was, indeed, used. Also shown are some security features in the actual sending of funds, including a thumb print scan and password. One would assume that Apple Pay will play a role in all this, and speculation has abounded that the new platform is meant to rival the Venmo system.

About a year ago, reports were in circulation that Apple and banks were discussing the implementation of such a system, and indeed this patent was first filed over a year ago. Presumably it will see ratification within the next year or two, supposing that other major AI developers like Microsoft choose not to contest it on some of its broader implications.

Microsoft and Google would surely oppose paying patent licensing fees to Apple in order to instantiate their own AI peer-to-peer payment systems, and for the most part, the patent application seems to indicate that Apple would then have sole dominion over that field. Therefore, with numerous dogs in the fight including Cortana, the future of this patent and/or the platform itself (failure to acquire a patent would not prohibit Apple from creating the platform, necessarily) is still very much up in the air. More is sure to surface in the coming months, and Hacked will keep you posted.

Images from Shutterstock and USPTO.

Important: Never invest money you can't afford to lose. Always do your own research and due diligence before placing a trade. Read our Terms & Conditions here.



Feedback or Requests?

Continue Reading

Trending