AI Develops its Own Language and Problems
February 25, 2019
The rapid pace of artificial intelligence development is not news. For years we have seen these rapid advances on the horizon, but only recently have we reached that tipping point. One of the most interesting recent areas of growth were language, which can become an issue for some professionals. We discovered more from the recent Psychology Today story, “Google and Facebook AI Make New Linguistics Discovery.”
According to the story:
“The researchers found that “the success rates between self-play and paired-play are indistinguishable from each other, strongly implying that a common, shared language emerges as a social convention if and only if we have more than two language users,” and all “that is needed in order for a common language to emerge is a minimum number of agents.”
It’s this independent development of language that has the intelligence community spooked. Some innovators are currently already at overload trying to detect invisible cyber threats.
Is this a problem that authentic AI can solve?
Patrick Roland, February 25, 2019
Amazonia for February 25, 2019
February 25, 2019
Several yellow flags flapped in the wind last week. In two conversations with conference organizers focused on the law enforcement and intelligence markets, I learned there was little interest in Amazon’s policeware services. I found this interesting but understandable. Amazon’s “footprint” is much larger in the eCommerce mindspace and recent news has been dominated by Amazon’s response to some New Yorkers’ protests over tax breaks for a cash rich, profitable company. Another factor is the ongoing background buzz about the suddenly personal life of Amazon’s founder. Nevertheless, DarkCyber Annex believes that Amazon is likely to be a disruptive force in what we call policeware and intelware. A few highlights from last week’s Amazon team:
Do Not Fear Amazon
“We’re from Amazon. We’re here to help you.” Jeff Bezos appears to be spending some cycle time promulgating these messages to employees. CNBC reported that “Jeff Bezos told employees that fear of Amazon is overblown.” According the CNBC:
Fears of Amazon taking over the world have reached a fever pitch in recent years. In the fourth quarter of 2017, Amazon was the most mentioned company on earnings calls of S&P 500 companies. But some of Amazon’s primary competitors are finding ways to survive and even thrive against one of the most valuable companies in the world.
Evidence is the positive financial performance of Wal-Mart and the number of misfires Amazon video has delivered. Also, Microsoft is catching up with Amazon cloud services. Rest easy.
Amazon Expands Its Threat Detection
There are many specialist companies — many of which will be acquired or just go out of business. Amazon is aware of this market. One possible reason is that many threat detection firms use Amazon’s infrastructure to provide their for fee services. It seems logical that Amazon would compete in this sector. InfoQ reported that Amazon had added three new threat detection services to its cyber security offerings. Amazon brands this initiative as GuardDuty. Infoq explains the service this way:
Amazon GuardDuty is a threat detection service available on AWS that continuously monitors for malicious or unauthorized behavior to help customers protect their AWS accounts and workloads. When a threat is detected, the service will send a detailed security alert to the GuardDuty console and AWS CloudWatch Events – thus making alerts actionable and easy to integrate into existing event management and workflow systems.
Amazon’s spin is that its existing customers can use these services. However, scope creep is likely to occur. Amazon may compete with some of its customers as it expands its revenue streams in this lucrative market.
Graphus Becomes AWS Partner
It’s difficult to keep track of the companies racing to become AWS partners. We noted that Virtual Strategy reported that Graphus is on the Bezos team. Graphus is a cyber security firm.
Ethereum Service Enhancement
Many individuals in government are not aware that Amazon is a player in the burgeoning digital currency game. Amazon is a player and an increasingly important one. Ethereum World News reported that Amazon supports deployment of VeChain Thor (VET) DApps with almost one click simplicity. What does this digital currency jargon mean? One one hand, an Amazon customer can deploy his or her own blockchain application without having to do bare metal coding. In terms of law enforcement, the expanding Ethereum services signal that data flowing into the through the Amazon system may well be of significance when it comes to identifying certain interesting behaviors. This development complements a managed blockchain service and a quantum ledger database.
Amazon Subsidies
A surprising subsidy from Smartronix reduces the cost of AWS cloud migrations has been reported by Globe Newswire. The idea is to reduce the cost of AWS migration for Virtual Machines (VM) running on VMware. For projects with a minimum of 500 VMs, the migration will be free (completely funded), with partial subsidies offered for smaller projects. In short, this is a play to get big installations. The idea is that organizations will be able to use the same tools and management capabilities they are using today, including VMware provisioning, storage, and lifecycle policies.
Amazon Snow Globe for Elastic Cloud Services
It is now possible to run an application within Amazon. The innovation is described in “Setting Up PrivateLink for Amazon ECS and Amazon ECR.” The idea is that “all the network traffic within the AWS network. When you create AWS PrivateLink endpoints for ECR and ECS, these service endpoints appear as elastic network interfaces with a private IP address in your VPC.” The idea is a variation on the catchphrase “What happens in Vegas, stays in Vegas. The Amazon edit becomes “What happens in AWS stays within AWS.”
Amazon Funds Computer Science Programs
Free training? Sounds like a promising offer. PC Magazine reports that Amazon will fund computing courses for under privileged teens. The online news service said:
More than 1,000 high schools across the US through its Future Engineer program [will receive funding]. Of those schools, more than 700 are classified as Title 1, meaning a high percentage of their students come from low-income families.
But anyone can become an Amazonia for Just head to Geek Deals and pay $35 for the AWS Certified Architect Developer Bundle 2019, now discounted by over 90 percent.
Artificial Intelligence: More Realism Is Not What Overvalued AI Firms Seek
February 24, 2019
I read an article which contained this statement about smart software:
What the AI actually did was urge judges to be more lenient with white people than black people.
Dangerous territory, right?
The source was “Artificial Intelligence: You Know It Isn’t Real, Yeah?”
The write up also points out:
AI is just a bunch of algorithms acting on data it is fed by its human programmers.
Where does this story go in terms of explaining the “reality” of smart software?
Bias is baked in.
Here’s an example from the write up:
Fair objectives are not enough if the data itself is biased. For example, if the courts have been over-sentencing black people for generations – as a result of ignorance, social conditions of certain racial groups or plain malice in the justice system, who knows – letting an AI rip on the unbalanced data simply trains it to be similarly biased. Hiding a field labeled “skin color” does not compensate for anything when the AI’s algorithms charge ahead identifying the same patterns of biased social profiling by the justice system anyway.
Interesting. But those valuations pivot on PR, hyperbole, unrealistic expectations, and the payoff from an exit strategy.
AI is another way for some to surf for a big payday. Accuracy? Reality? Authentic AI? Not necessary.
Stephen E Arnold, February 24, 2019
Less Fear? More Benefits?
February 23, 2019
Our world has never been more technologically advanced, that’s a fact. That also means that the digital threats have never been more dire, right? Yes and no, according to one source, who says that the technology might change but humans never do. We learned more from a recent CNBC story, “Google Infosec Head Heather Adkins: Ignore Scare Stories.”
According to the story:
“Adkins said sometimes the marketplace suffers from a “proliferation of cybersecurity professionals” offering conflicting advice on passwords, antivirus software, safety practices and so on…But the best rules for individuals looking to secure their personal information are the classics, Adkins said…Keep your software up to date, and don’t re-use the same password.”
This and many other examples show that good old fashioned foresight and detective work can still help fight cybercrime, even in this world of machine learning and nanotech. Now is a good time to rethink how savvy cyber organizations think about security, as the Monica Witt hacking story proves, sometimes we’re over thinking things and leaving the back door wide open. As Adkins says, let’s look forward in regards to security, but also not forget our past.
Patrick Roland, February 23, 2019
Google Smart Software: Yeah AI Is Super
February 22, 2019
DarkCyber did notice the flurry of write ups about kid vids and the “comments” once available on YouTube. I am an adviser to an international tribunal focused on human trafficking and sex crimes. Suddenly quite a few “Web search experts” and “YouTube media consumers” have discovered interesting content tucked into the YouTube service.
I found this headline representative. It is from CNBC, a real news outfit; to wit:
As Fallout Over Pedophilia Content on YouTube Continues, AT&T and Hasbro Pull All Advertisements
I like the use of pedophilia in a mainstream real news story. I like the certainty of “all”.
The problem is that the content is available on YouTube and that the comments sections for videos harbor a wide range of content some may find objectionable. Need to hack commercial software. There are videos which include the how to as text over a video about not much at all. Weapons? Backyard detonations? Substances to ingest when happy or sad? And more.
The question I have is, “Why are videos with questionable material accepted by YouTube?”
When I post one of my videos, I have to log in. I have to fill out a form. The system “selects” an image for my video. I have to put in key words. I have the option to describe my video.
With the hoo hah about Google’s DeepMind and the artificial intelligence which does the work of matching images with content, why are child related videos and comments allowed to slide through.
One of my researchers found a video on YouTube which goes through the procedure for locating kiddie imagery. I won’t put the link in this blog, but a how to about this type of content? Come on, man, as the talking heads on football programs have been known to say. Come on, man.
Now about the pulling of advertising. Since big companies want to sell to those who watch YouTube videos, my hunch is that most of the outraged commercial enterprises will return to YouTube. Perhaps one of these companies will advertise 5G connectivity when it is not yet available.
Come on, man.
Whether it is Facebook’s Onavo, Microsoft’s app store harboring crypto mining malware, or Apple putting servers in Russia — there seems to be a common thread. I see these examples as useful insights into what the high school science club approach to management fosters. Heh, heh, heh, we’re really sorry.
Yep, I believe that.
Look at those ethical compasses. Spinning like crazy and powering the machines that are distinctively Silicon Valley.
Stephen E Arnold, February 22, 2019
Predicting Human Nature With AI1
February 22, 2019
A potential leap forward in predicting human behavior through AI is originating in a very unlikely place: The spice aisle of your grocery store. Seriously. AI and data analytics are moving into interesting territory and the results might be of interest to as far reaching places as the intelligence community. We learned more from a recent Engadget story, “How McCormick and IBM Will Use AI to Create the Next Big Spice.”
According to the story:
“IBM Research unveiled a version of this technology the Philyra AI as a tool to accelerate the creation of new and novel scents for the fragrance industry. “It is a system that uses new and advanced machine learning algorithms to sift through hundreds of thousands of formulas and thousands of raw materials…helping identify patterns and novel combinations.”
While this might seem silly to pay attention to, the results could be a bigger sign than simply what we should be sprinkling on our dinner. In fact, if this experiment does yield something customers truly want, it’s a sign of AI’s ability to predict human nature. This doesn’t necessarily make for happy bedfellows, as evidenced by the recent uproar of Palantir partnering with intelligence organizations. Keep an eye on these developments, because we’ll be feeling the aftershocks for years.
And let’s use these tools to spot objectionable content.
Patrick Roland, February 22, 2019
Dark Web Directory: Updates Needed
February 22, 2019
If the Internet were an ocean, the Dark Web is a very shallow tide pool. While the Dark Web is shallow, we do not recommend diving in because you can still break your neck. The Dark Web has a limited number of Web sites listed on it, all of them using the .onion extension.
These Web sites are accessible using the Tor browser and you do not use a search engine to find them. Instead you rely on social media Web sites, such as reddit, forums, or the Dark Web News. The Dark Web News has the “Dark Web & Deep Web Market List With Up & Down Daily Updated Market Status.”
The market listing is described as “Are you wondering how to find deep net markets? Well, look no further! We have compiled a list of active hidden marketplaces available on the deep web.” It is followed by a guide on how to access the Dark Web, download the Tor browser, etc.
What is striking is the amount of warnings about losing your anonymity. The market listing states, no shouts, that a smart Dark Web user uses not only the Tor browser, but also has a VPN to encrypt their data.
After the anonymity warnings, there are the Dark Web market listings. Each market site is reviewed, given a small description, and its status is shared. The listings are very useful and help track the type of market you are searching for. The only downside is that it lists Silk Road and a few other places as still “open.” Methinks that the Dark Web market listing needs an update. Also they give another good warning: “Do your research before using any hidden marketplace. Reddit is a good place to start.”
The problem is that the Dark Web is not zipping along as it once was. The buying and selling action has shifted to online chat and closed discussion groups. As the Dark Web shrinks, maintaining a listing should be easier too.
Whitney Grace, February 22, 2019
Google Adds to Its Fancy Dancing Repertoire
February 21, 2019
I assume that someone at Google learned about the UK report hashtagging Facebook as a “digital gangster.”
Google is almost certainly aware that regulatory scrutiny of the firm’s practices is likely to increase in 2019. One of Google’s easier dance moves is reported in “Google Exec Reorganizes Policy Shop as Global Threats Loom.” Nothing solves problems like influencers, insiders and money. The write up asserted:
“Public Policy,” will become “Government Affairs and Public Policy.”
Ah, wordsmithing.
Tougher to explain is the report is another wave of advertisers (many of which have no other way to promote their products and services) are finding themselves taking a mor-tical stand. (That’s a combo of moral and ethical, a neologism for online marketing.)
I noted “Nestle, Disney Pull YouTube Ads, Joining Furor Over Child Videos.” The sometimes source free Bloomberg reports:
Walt Disney Co. is said to have pulled its advertising spending from YouTube, joining other companies including Nestle SA, after a blogger detailed how comments on Google’s video site were being used to facilitate a “soft-core pedophilia ring.” Some of the videos involved ran next to ads placed by Disney and Nestle.
Bloomberg, true to real news norms, adds this statement:
YouTube on Tuesday released an updated policy about how it will handle content that “crosses the line” of appropriateness.
I don’t want to dwell on appropriateness, lobbying, or the cycle of surprise, apologies, and remediation which seems more like a visit to the previously owned and lightly used shop.
Like Facebook, Google has some interesting methods of generating revenue as it continues to avoid the “digital gangster” moniker. Inappropriate kiddie content is, however, problematic for any organization. How? Why? How much? Who? — Questions which may warrant answers some day. Maybe.
It may be time for the founders to distance themselves even more from the online ad giant. The quite valuable 25 year old teapot may be reaching it limit for safe operation.
Stephen E Arnold, February 21, 2019
AI: The Facebook View for the Moment
February 21, 2019
We get some insight into the current trajectory of AI from Fortune’s article, “Facebook’s Chief A.I. Scientist Yann LeCun On the Future of Computer Chips, Lawnmowers, and Deep Learning.” The write-up points to a talk on AI hardware LeCun gave at the recent International Solid-State Circuits Conference in San Francisco.
Writer Jonathan Vanian highlights three points. First, he notes the advent of specialized chips designed to save energy, which should facilitate the use of more neural networks within data centers. This could mean faster speech translation, for example, or more effective image analysis. The tech could even improve content moderation, a subject much on Facebook’s mind right now. Then there are our “smart” devices, which can be expected to grow more clever as their chips get smaller. For instance, Vanian envisions a lawn mower that could identify and pull weeds. He notes, though, that battery capacity is another conundrum altogether.
Finally, we come to the curious issue of “common sense”—so far, AIs tend to fall far short of humans in that area. We’re told:
“Despite advances in deep learning, computers still lack common sense. They would need to review thousands of images of an elephant to independently identify them in other photos. In contrast, children quickly recognize elephants because they have a basic understanding about the animals. If challenged, they can extrapolate that an elephant is merely a different kind of animal—albeit a really big one. LeCun believes that new kinds of neural networks will eventually be developed that gain common sense by sifting through a smorgasbord of data. It would be akin to teaching the technology basic facts that it can later reference, like an encyclopedia. AI practitioners could then refine these neural networks by further training them to recognize and carry out more advanced tasks than modern versions.”
The chips to facilitate that leap are not yet on the market, of course. However, LeCun seems to believe they will soon be upon us. I do hope so; perhaps these super chips will bring some much needed sense to our online discourse.
Cynthia Murrell, February 21, 2019
Regulation: Social Media Is Not a Right
February 20, 2019
Will this ruling push more social-media bad actors onto encrypted communications? Nextgov reports, “A California Court Finds Social Media Posts Aren’t a First Amendment Right.” The headline is a tad inaccurate, I’m afraid. The appeals court ruling actually places a rational limit on that Supreme-Court affirmed right (think of the “fire in a crowded theatre” example.) In fact, the case involved a juvenile on probation who was apparently inclined to brag about his crime. Reporter Ephrat Livni writes:
“Basically, it’s true that AA has the right to speak freely. But his freedom is legally curtailed by probation conditions designed to rehabilitate him and protect the victim. The court noted that his social media posts could endanger the victim in this case and that there were plenty of other ways AA could communicate, including email, phone, in person, and via written correspondence. To the extent that AA’s case seems to contradict the conclusions about free speech and social media in Packingham, it’s notable that the Supreme Court case involved a North Carolina law that made it a felony for sex offenders to have any social media presence at all, indefinitely. In this juvenile case, however, AA was only barred from posting about his offense and only for the duration of his probation.”
So the perpetual dance between freedom and protection continues. Livni also notes that “AA” claimed his attorney was at fault for failing to object to the prohibition in the first place, but that plea was dismissed out of hand. So, freedom of speech on social media is indeed a First Amendment right (at least for now), but, as always, common sense must be applied.
Cynthia Murrell, February 20, 2019