GENERATION GENERATION
the output has a defined form, the more data provided in the request, the more specific the result. Te output is generally not to the point where it can be deployed unsupervised, but the models can provide a useful first draft for a human to review and revise.
One of the most impressive capabilities of generative pre-trained models is their ability to respond accurately to prompts. However, domain-specific applications, like Army operations, may require post-training of these models for more accurate responses. Beaudry explains that “even within the green-suit Army, we speak different languages. If I, as an artilleryman, were to say that our objective was to ‘destroy the target,’ that creates a differ- ent impression in the minds of an infantryman or armor Soldier than it does to artillery, where artillery destruction is defined as one-third destroyed versus destroying everything using the infan- try and armor mindset.”
TELL US HOW YOU FEEL
Generative AI’s ability to simulate multimodal sensory information in a virtual environment can help teach Soldiers how to identify objects, like how a potato feels or how half-buried anti-tank mines look, without a hazardous learning environment. (Image generated by DALL-E 2)
Command Battle Lab, notes that his organization is looking into. Te design can move beyond the planning phase and into problem solving during operations, which allows autonomous action in pursuit of a delegated objective.
OFFLOADING BUREAUCRACY While applications like DALL-E 2 and Stable Diffusion have popularized AI-generated images on social media, OpenAI’s generative pre-trained models, based on massive amounts of text from public documents and the internet, have both impressed and horrified users with their capability.
Much of the focus for generative AI has been focused on the negative impacts—the end of the student essay, the loss of jobs for professional illustrators—but there are also positive impacts.
Generative AI can be used to develop document outlines (see “Prompt” box), draft sections of documents (such as a concept of operations) or correspondence, write abstracts or summarize documents, and even write code, reducing the amount of time and effort involved. While the AI works best on requests where
32 Army AL&T Magazine Spring 2023
OpenAI has at least mitigated one of the early problems with generative text models. When a model is trained on such massive amounts of data, it becomes difficult to control the data for qual- ity. Early models betrayed the ignorance of trolls in the data they consumed, producing, at times, horrifying output. Since that time, content guardrails (ChatGPT will refuse to write an ode to your armpit, but an earlobe is fine) as well as “charm school” training, with evaluators rating the model’s responses, have made these tools viable for general use.
CONCLUSION Looking forward, the potential for generative AI to be used by the Army, in both operations and for enterprise use, has clear benefits for transformation—in the hands of those with the experience, expertise and willingness to embrace new ways of working. Tat will make the potatoes feel great.
For more information, contact Thom Hawkins
jeffrey.t.hawkins10.civ@
army.mil.
THOM HAWKINS is a project officer for artificial intelligence and data strategy with Project Manager Mission Command, assigned to the Program Executive Office for Command, Control and Communications – Tactical, at Aberdeen Proving Ground, Maryland. He holds an M.S. in library and information science from Drexel University and a B.A. in English from Washington College. He won an ALTie award for his article, “Outside the Lines” in 2014.
at
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62 |
Page 63 |
Page 64 |
Page 65 |
Page 66 |
Page 67 |
Page 68 |
Page 69 |
Page 70 |
Page 71 |
Page 72 |
Page 73 |
Page 74 |
Page 75 |
Page 76 |
Page 77 |
Page 78 |
Page 79 |
Page 80 |
Page 81 |
Page 82 |
Page 83 |
Page 84 |
Page 85 |
Page 86 |
Page 87 |
Page 88 |
Page 89 |
Page 90 |
Page 91 |
Page 92 |
Page 93 |
Page 94 |
Page 95 |
Page 96 |
Page 97 |
Page 98 |
Page 99 |
Page 100 |
Page 101 |
Page 102 |
Page 103 |
Page 104 |
Page 105 |
Page 106 |
Page 107 |
Page 108 |
Page 109 |
Page 110 |
Page 111 |
Page 112 |
Page 113 |
Page 114 |
Page 115 |
Page 116 |
Page 117 |
Page 118 |
Page 119 |
Page 120 |
Page 121 |
Page 122 |
Page 123 |
Page 124 |
Page 125 |
Page 126 |
Page 127 |
Page 128 |
Page 129 |
Page 130 |
Page 131 |
Page 132
