On Thursday, Google released a document entitled "Artificial Intelligence at Google: Our Principles," vowing to avoid Pentagon projects to develop AI weapons. However, they'll continue to work with the US military on a host of other projects, including AI projects, so long as they don't include surveillance that runs counter to human rights.
"Google is in a spot at the moment," Wallis, editor-at-large for Digital Journal and author of more than a dozen books, told Loud & Clear hosts John Kiriakou and Nicole Roussell. "This is gold rush time for artificial intelligence; everybody wants a piece of it, and we're not talking about the sort of generic type of artificial intelligence — we're talking about possibly thousands or millions of different kinds of artificial intelligence."
In other words, the advent of artificially intelligent weapons is "inevitable," Wallis said.
"Google has been working on core systems like autonomous vehicles and so forth, no doubt with the best of intentions, but it doesn't follow that that class of technology can't be used as a sort of spring board for autonomous weapons systems and so forth," Wallis said. "The whole history of military development since the bronze age has been highly competitive."
Wallis elaborated further: "If you think about it, the original tank was developed from farm equipment. Now, imagine a fridge which can order your groceries and simultaneously fire off a few hellfire missiles when it feels like it and remotely control a UAV [unmanned aerial vehicle, or drone] somewhere over the Middle East or wherever. The internet of things could turn into a terrible own goal in that regard: it can never be secured, it can never be properly monitored at any given time. It's a sort of accident waiting to happen, but the thing is that it's going to be using much the same technology and you can sort of patch it through to whatever system it needs to run."
Wallis spoke of an AI arms race to come, telling Loud & Clear that there are "God knows how many forms of corporate entities trying to compete and get into what is going to be the hottest property in artificial intelligence." And so, "what Google does or doesn't do is ultimately going to be driven by the market," he said.
The Defense Advanced Research Projects Agency (DARPA) could circumvent Google's protest by developing its own AI or by seeking another contractor "with ease," Wallis said.
"DARPA is in a very tough position at the moment," he noted, adding that "the current state of technology is that the next generation of systems will be artificially intelligent. The problem with that is that their choices are: they don't develop the technology, or, let the other guys develop the technology. From a purely military point of view, there's no reason believe that other countries will not develop artificial intelligence."
At the end of the day, the US military, no matter how much more advanced it gets than its competitors, "can't really dodge that. So whether Google's involved or not is going to be a relatively minor issue," he said.
For now, he suspects Google's new policy is only temporary. "My feeling is that Google will have to be involved at some point. I think DARPA would prefer if Google was involved but the fact is the market is going to drag it along whether it likes it or not."