US Air Force Wants to Use AI Technology to Gather Intelligence From Social Media

US Air Force Chief of Staff General David Goldfein has expressed the force's desire to obtain major artificial intelligence (AI) computing capabilities to help them better gather intelligence. That the Air Force wants to gather intelligence is not surprising: what is, however, is that Goldfein said that they want to find it on social media, and primarily from Twitter.

To illustrate his position, he used the example of the MH17 plane crash.

"When the Russians shot down the airliner, and we were searching for the smoking gun, we found it a month later — on Facebook," the general said at an Air Force Association breakfast in Washington Wednesday, according to DefenseTech.com. Goldfein pointed the finger at Russia for the crash, as though the Netherlands have already made a conclusion as to who shot the Buk missile that brought the plane down (which they have not).

"We found posted pictures on Russian blog sites that actually showed the activity, but it took us a month to figure that out," Goldfein said, leaning on social media as a source of reliable information, even though investigation of the MH17 catastrophe is still going on and it is hampered by "a great deal of disinformation and attempts to discredit the investigation," Dutch Foreign Minister Bert Koenders said in a statement earlier in July, according to France 24.

Social media is arguably the least moderated print information source in the world; in fact, it is so riddled with deliberately engineered misinformation that Facebook even had to issue tips to help users discern between truth and lies. The company even considered hiring an army of paid fact-checkers in an effort to reduce the flow of fake messages.

Discussing a trip to the offices of the Bloomberg news agency, he reportedly asked a technician to perform a Twitter search on violent extremist activity over the last 48 hours, and the system actually mapped the relevant tweets on a map.

"I watched as a map populated with little circles… All of it from real-time analysis from Twitter with corresponding values associated with each account… related to the ability to determine if this Twitter report is mirrored by others," Goldfein said, according to International Business Times.

However, what Goldfein wants is for a machine not just to map tweets, but to conduct analysis, and to provide refined information to the operator in a fast and understandable manner. This raises concerns that the AI might take over the human role of understanding the data, and whether algorithms can be trusted to make the right conclusions from the data being fed into the machine.

"Goldfein on Wednesday reiterated that, like the Bloomberg tech, ‘refining the analysis' through AI to reach the operator, pilot or airman in a decisive and streamlined way is what the Air Force desperately needs to conduct missions in the future," DefenseTech reports.

However, the mutually exclusive information that Facebook and other social media platforms frequently contain would seem more likely to confuse than assist operators, pilots and airmen.