deskilling of the workforce until a failure occurs.
Kevin Burns (MITRE) presented the results of a relia-bility model showing the potential safety impact of a
cognitive copilot and what features would provide
the most benefit.
The session on cybersecurity included examples by
Gheorghe Tecuci (George Mason University) and Jim
Whitmore (formerly of IBM), followed by a lively discussion concluding that the volume of cyber attacks
could force the move toward automation even if cogs
don’t perform equal to human experts.
There was a session on natural language processing for US federal acquisition regulations for adjudicating administrative law and one on using NLP to
monitor social media for situational understanding.
We also had presentations on cognitive assistance for
dealing with autism, helping with housing for homeless youth, and education.
The symposium audience asked if AI can be used to
generate fake images, videos, and blogs, how will we
know truth from fiction? If AI identifies a new drug,
who gets the credit? If AI misidentifies a disease, who
gets the blame? How do we adapt the workforce to
working with cognitive assistants as collaborators
and partners? The participants agreed that they
would like to attend future symposia to share experiences and address some of the challenges posed this
Frank Stein (IBM) and Chuck Howell (MITRE)
served as cochairs of this symposium. The organizing
committee included Lashon Booker (MITRE), Chris
Codella (IBM), Eduard Hovy (Carnegie Mellon University), Anupam Joshi (University of Maryland, Baltimore County), Andrew Lacher (MITRE), Jim
Spohrer (IBM), and John Tyler (IBM). The session
papers were published as AAAI Press Technical Report
Deep Models and Artificial
Intelligence for Military
Applications: Potentials, Theories,
Recent advancements in artificial intelligence enable
new technologies to assist modern warfighters by
automatically analyzing big data at timescales much
faster than a human can achieve. Deep learning (DL)
is the core of the new AI revolution, demonstrating
that not only can machines classify quicker than
humans, but they can also classify more accurately
than humans. These technologies have revolution-
ized many commercial applications but are not
designed to solve military problems.