Maven Initiative: More AI for US Military?
May 2, 2018
We receive an update on AI in the field of intelligence from International Business Times in their article, “What Is Project Maven? The Pentagon’s New AI Is Hunting Terrorists in the Wild.” Machine-learning tool Maven was developed in-house, and has already been put into service in the Middle East. As good AI’s are wont to do, we’re told the platform learns from experience and has reached an 80% accuracy rate in target identification. That sounds pretty good. (Unless one finds oneself in the 20%, I suppose.) Paired with drones, Maven can be quite effective; it is designed to interface with the georegistration system Minotaur, currently in use by our Navy and Marine Corps. Reporter India Ashok writes:
“While the AI system was deployed in the Middle East, it reportedly helped US intelligence analysts identify objects in a video of a battlefield captured by a ScanEagle drone. In 2018, the Project Maven team will reportedly work on automating battlefield drone video analysis. By the next summer, the team reportedly hopes to incorporate the AI system’s capabilities with larger UAVs such as the Predator and Reaper drones. The team is also reportedly planning to incorporate Maven into the Gorgon Stare, a high-tech series of cameras mounted on Reaper drones, which can view entire towns. Equipping such Reaper drones with the Maven algorithm will likely allow intel analysts to cast a wider net around the area and also likely identify civilian areas. Maven’s deployment might not only revolutionize war tech, but also reshape the way intel analysts perform. The success of the project might also likely boost support for AI operations further in other defense programs.”
Air Force Lt. Gen. John N.T. Shanahan sees the project as the beginning of an effective partnership between humans and machines. He believes every aspect of the Department of Defense could be enhanced by machine learning and expects that, by the end of the year, analysts will have developed many creative uses for the ever-evolving Maven. No doubt.
Cynthia Murrell, May 2, 2018
Former Autonomy Executive Found Culpable
May 1, 2018
I read “U.S. Jury Convicts Former Autonomy Executive of Fraud over HP Deal.” The jury found Sushovan Hussain guilty of wire fraud. Reuters said that in 2009 Mr. Hussain began to “deceive Autonomy’s investors and HP( about the company’s financial condition and prospects for growth.”
An ArnoldIT profile about Autonomy is available without charge at this link.
The HP purchase of Autonomy was one of the major turning points in enterprise search. The $11 billion deal made clear that enterprise search was an application space which would have provided rocket fuel for HP’s growth.
HP discovered that enterprise search was a challenging technology to use as a way to generates billions of dollars in revenue quickly and easily.
Beyond Search’s view of this deal—as well as the sale of Exalead to Dassault Systèmes, Vivisimo’s sale to IBM, and the manic repositioning of the vendors pitching proprietary search solutions—is that may companies found that enterprise search was a tough nut to crack. Marketing is easier than generating sustainable revenues. In my experience, enterprise search requires specialized expertise.
What’s interesting is that I heard that HP executives reviewed the Autonomy financial data, knew about the Qatalyst 2011 report about Autonomy, and decided to purchase the company. The deal seemed to unfold quickly and then implode almost as quickly. HP paid more for Autonomy than any other search acquisition of which I was aware. HP emerged as the proud owner of the firm which brought Bayesian methods to the enterprise. (I want to mention that the mathematical procedures implemented by Autonomy are now incorporated into most of the next generation information access systems I discussed in CyberOSINT, my book on what’s beyond search.
A page from the 2011 Qatalyst report about Autonomy. The full document, once available on the Oracle Web site, is now difficult to locate via open source methods.
This legal dust up may gather momentum. Possibly appeals and more people accused by HP making their way to the court room. I will stay tuned for developments reported by “real news” outfits like Thomson Reuters.
Stephen E Arnold, May 1, 2018
Geotargeting: Getting Popular
May 1, 2018
Businesses, governments, and organizations are asking, “How can AI be used?” The better question to ask is, “What can AI not do?” Along with spying on the Chinese’s good behavior and people’s personal information on social media, the Smart Data Collective posted that “Malicious AI? Report Shines Dark Light On Geotargeting.”
What is geotargeting? Geotargeting is using locations specific data to keep a close eye on selected targets. Social media data plays a part too. It is astonishing and creepy how much AI can pull from information placed on the Internet. Ever since Russia intervened in the 2016 election, policymakers are cracking down on data-based marketing. What is even worse is that hackers are already using AI against the innocent.
New policies are being put into place and Mark Zuckerberg is even being held (hopefully) accountable for how Facebook has taken advantage of data.
“People are growing more and more suspicious of AI. The new malicious AI report, which was written by 26 experts from academia, industry, and 12 other fields, identifies AI as a potential culprit in the threat to “political security.” The report says that AI “can automate tasks involved in surveillance” by analyzing “mass-collected data,” which it can use to create propaganda and deceptive content, such as misleading videos and fake news. The more trolls and hackers use AI to threaten the political security of democracies, the more likely democracies and companies are to regulate the use of big data.”
One scary AI trick is geotargeting, where companies can push advertising directly to customers’ and law enforcement can track people of interested all based on a persons’ cell phone data. Bad actors can use geotargeting for personal drone attacks. Scary! There are not any amendments currently protecting AI, so the legal ramifications are still up in the air.
Remember, however, that AI is a neutral tool. Humans are creatures of habit. Plotting geo-behaviors can be a useful, insightful exercise.
Whitney Grace, May 1, 2018
DarkCyber for May 1, 2018, Now Available
May 1, 2018
DarkCyber is a weekly video news program which covers important Dark Web stories and information about less well known Internet services. Produced by Stephen E Arnold, publisher of the Beyond Search blog, DarkCyber is available at www.arnoldit.com/wordpress and streaming on Vimeo at https://vimeo.com/267103171 .
Russia has blocked Telegram, the popular messaging app which had an estimated nine million users in Russia. DarkCyber explains that Russian government officials must now use decades old technology for their text messages. One consequence of the Russian blocking of Telegram is that service to Amazon and Google was interrupted. DarkCyber provides a workaround that Russian users may want to consider adopting to respond to the stepped up censorship in Russia.
A new report from a unit of the GHCQ (Britain’s equivalent of the US National Security Agency) provides a thorough run down of cyber crime activity in England. DarkCyber highlights how a person can download a free copy of this important report. Plus, DarkCyber describes a case example of Crime as a Service highlighted in the study. The particular CaaS involves an individual providing malware programmers a way to verify that their code could elude some detection systems. Plus, DarkCyber reveals how the bad actor provided his paying customers with free customer support.
DarkCyber provides basic information explaining how a person can set up a Dark Web server. The procedure is straightforward but may be too complex or cumbersome for some users who want to take advantage of Tor’s anonymity features. DarkCyber provides an easy solution which can get a Dark Web site online in a matter of minutes and costs pennies a day.
The final story reiterates a theme based on a person’s assumption that the Dark Web is anonymous. For an individual who believed that Ecstasy purchases with payment via Bitcoin were invisible to law enforcement, the Dark Web is not as Dark as she assumed. Australian and UK authorities arrested the person who assumed incorrectly that Tor was 100 percent anonymous.
We have also updated Stephen’s brief biography. We have reproduced it below:
Stephen E Arnold is the author of “Dark Web Notebook” and “CyberOSINT: Next Generation Information Access.” This book describes some of the technologies used by GSR and Cambridge Analytica to acquire and analyze Facebook user data. He has been named as a technology adviser to the UK based Judicial Commission of Inquiry into Human Trafficking and Child Sex Abuse.” Mr. Arnold also lectures to law enforcement and intelligence professionals attending the Telestrategies ISS conferences in Prague, At that conference, he will describe a major vendor’s virtually-unknown digital currency deanonymizing service. In addition, Mr. Arnold will appear at the Washington, DC, and Panama City, Panama, Telestrategies ISS events. In recent months, he has shared his research with law enforcement and intelligence professionals in the US and Europe. His most recent lectures focus on deanonymizing chat and digital currency transactions. One hour and full day programs are available via webinars and on-site presentations. He publishes the free Web log “Beyond Search,” which is available at www.arnoldit.com/wordpress .
DarkCyber is available at this link. (The splash page for the video contains a nod to May Day celebrations in a certain country.) We are working on a special DarkCyber about Amazon’s “intel play” which will be released coincident with his lectures at the Telestrategies ISS conference in Prague during the first week of June.
Kenny Toth, May 1, 2018
-
- Subscribe to Beyond Search
Feature archive
News archive
-
