The AI arms race continues: AI and machine learning startups have been popping up like mad in recent times, and are being acquired just as quickly. This week, Facebook and SaaS application vendor LogMeIn continued the trend, buying Ozlo and Nanorep, respectively.
Ozlo's capabilities will be rolled into Facebook Messenger to help it build "compelling experiences" that are "powered by artificial intelligence and machine learning," Facebook said. For its part, Ozlo says it has built a knowledge graph that contains more than 2 billion entities, making its technology able to "understand real-world nuances." Translation: Facebook wants to improve its chatbots level of understanding and interaction with users. Terms of the Ozlo deal weren't disclosed.
Meanwhile, LogMeIn is paying about $50 million for Nanorep, an Israeli startup focused on customer self-service, chatbot and virtual assistant technology. The deal will complement LogMeIn's recently launched customer engagement platform, Bold360.
POV: The flurry of AI acquisitions is no surprise given the energy in the market, as larger vendors scramble to get their AI strategies in shape. It's telling that even a company with as many resources as Facebook is going down this path, suggesting that acquiring AI talent as much as technology is a key driver for these deals.
Bitcoin's big split: The world's leading cryptocurrency was forked on Tuesday by a minority group of developers backing a new iteration, which will be called Bitcoin Cash. The split comes after a couple of years of debate in the Bitcoin community over the cryptocurrency's blockchain architecture, which at present allows fewer than 10 transactions per second. Bitcoin Cash's approach will increase the currency's scalability.
The new currency faces some challenges, chief among them the fact that only some Bitcoin exchanges will support it for now. It also won't be able to take advantage of the services ecosystem built up around Bitcoin. In turn, the currency could cause market confusion among would-be investors, as well as cause headaches for companies building their businesses on Bitcoin.
POV: It will be some time before Bitcoin Cash's success or failure can be measured, as well as its impact on Bitcoin. On a philosophical level, there are strong arguments for and against it from an open-source software perspective. In the first case, the ability to fork code is a bedrock tenet of open source. But as one critic points out, Bitcoin Cash's technical approach could make its blockchain too large for smaller organizations to store in its entirety, leaving more control in the hands of fewer players—the exact opposite of the decentralized structure touted by Bitcoin from the beginning.
Malware via machine learning: The prospect of AI becoming a threat to humans is a popular topic, and new software created by security vendor Endgame will do little to quell that discussion. Endgame used the AI framework provided by Elon Musk-backed nonprofit OpenAI to generate malware that can defeat antivirus software, as the Register reports:
In a keynote demonstration at the DEF CON hacking convention Hyrum Anderson, technical director of data science at security shop Endgame ... cited research by Google and others to show how changing just a few pixels in an image can cause classification software to mistake a bus for an ostrich.
“All machine learning models have blind spots,” he said. “Depending on how much knowledge a hacker has they can be convenient to exploit.”
So the team built a fairly simple mechanism to develop weaponised code by making very small changes to malware and firing these variants at an antivirus file scanner. By monitoring the response from the engine they were able to make lots of tiny tweaks that proved very effective at crafting software nasties that could evade security sensors.
Endgame's software managed to defeat the security software 16 percent of the time, according to the Register.
POV: Security vendors already widely use machine learning for defensive means, and tools such as Endgame's should prove useful for penetration testing. And Endgame wasn't the only vendor at DEF CON demonstrating an AI-based hacking tool. Security consultancy Bishop Fox showcased DeepHack, which uses AI to break into web applications. Here's how it describes the approach:
DeepHack works the following way: Neural networks used in reinforcement learning excel at finding solutions to games. By describing a problem as a "game" with winners, losers, points, objectives, and actions, a neural network can be trained to be proficient at "playing" it. The AI is rewarded every time it sends a request to gain new information about the target system, thereby discovering what types of requests lead to that information.
While Endgame and Bishop Fox's intentions are benevolent, it's an open question as to how much AI becomes a tool for malicious hackers going forward.
