Could Do-It-Yourself AI Make People More Comfortable with It?


As artificial intelligence makes strides in government, it could be helpful to see for yourself what the tech can do.

Artificial intelligence, sometimes just called AI, has been at the forefront of a lot of people’s minds for some time now. Many of us have probably seen a few snippets of what AI can do, whether it’s IBM’s Watson helping to analyze our taxes, a self-driving vehicle (hopefully) safely passing us on the road, our favorite video games getting a lot smarter, or our phones “helpfully” offering us ads for services that it knows we will like.

Despite AI touching more of our lives, it’s still a technology that’s under development in a lot of ways. And quite a few very smart people feel it may be the end of us all. The most recent season of “The X-Files” tapped into a bit of that paranoia in an episode called Rm9sbG93ZXJz, which translates to “followers” in Base64, where an army of intelligent robots try to force Mulder and Scully to leave a tip for bad sushi.

Aggressive chef robots aside, the government is increasingly interested in AI. In May, the White House held a large AI Summit that brought leaders in business, research laboratories and academia together with government officials to talk about AI development, how AI could change the workforce and the economy, and what regulatory restrictions should be put in place to ensure that it all goes well.

In many corners of government, AI is already deployed. NASA’s Transiting Exoplanet Survey Satellite was launched this spring with more AI than almost anything we have ever put in space. Its mission is to search for extraterrestrial and intelligent life outside our solar system. That’s kind of ironic that our created intelligence is looking for alien intelligence on our behalf.

Of course, the military is also interested because using AI in combat could mean putting fewer humans in harm’s way. The Defense Department founded the Joint Artificial Intelligence Center in June. The center is tasked with initially overseeing 600 military projects involving AI in some form, and Defense hopes to add more to that roster soon. Meanwhile, the Defense Advanced Research Projects Agency is also investing in the science, allocating $2 billion for its AI Next project, which aims to advance AI technology into previously untouched areas.

It may seem like the government is going a little crazy on AI development, but this is a fascinating science. I admit that it has captured my imagination as well. That is why I was really pleased when I got the opportunity to review the Senzing AI program for CSO magazine. I have evaluated software in the past with AI elements, which can be particularly useful in cybersecurity, but this was the first time I got to test a program with pure AI at its heart.

While the Senzing program, which was spun off from an IBM project, won’t look for extraterrestrial life or drive your car, what it can accomplish is pretty fascinating. It’s designed to make connections between elements in any number of databases that no human probably could. It’s even able to compensate for things like typos, people using nicknames, data that was incorrectly entered and even people who lied and gave purposely false information. That is actually surprisingly powerful, especially if clever humans think of ways to gain insight from previously unusable or overlooked data.

Senzing is also giving away a lot of functionality in order to show people the benefits of AI. Anyone can download the entire program for free for either Mac or Windows, and it will run on any moderately powerful desktop system. It’s not connected to the cloud or anything like that, so you can even run it on an air-gapped network. Once you have the program, you can task it with examining and comparing up to 10,000 records for absolutely any purpose you want for no charge. And if you have a few billion records to scan, you can do that as well for the reasonable price of $55,000 per month.

In my testing, I loaded up a few thousand records from five databases, which was a pretty limited sample size but was still amazed at what the AI could accomplish. It was able to show me hidden connections between records in multiple databases based on the tiniest bit of information, even pointing out the ways that seemingly unconnected people had relationships with specific targets that I designated. It was almost like playing Six Degrees of Separation at warp speed, and even including people who would not ordinarily be located by traditional searching.

And because this is a true AI, it knew that new information might change its previous assumptions, something that even humans sometimes overlook. Every time I added a new database for consideration, the Senzing AI went back and double checked all of its previous findings. It did that to see if it could make any new connections, but also to see if any previous assumptions might be proven wrong when examined in context with the new data.

I was pretty pleased with my first foray into the world of AI and can see why the government is investing billions in researching every possible facet of it. I still don’t think that AI will ever pose a threat to humanity. Well-designed AI is sort of like someone with Savant syndrome in that they possess abilities far above most humans, but for very limited tasks. An AI might be able to make connections and provide insights better than any human, or even teams of humans, but doesn’t really care who you vote for, what you believe, or if you leave a tip at your favorite automated sushi joint.

John Breeden II is an award-winning journalist and reviewer with over 20 years of experience covering technology. He is the CEO of the Tech Writers Bureau, a group that creates technological thought leadership content for organizations of all sizes. Twitter: @LabGuys