Alex Alston's AI Statement:
"These models are fundamentally built by and on plagiarism and the theft of intellectual property to enrich a few oligarchs while destroying the environment and public capacity for cultural memory (they literally destroy old books). This destruction is socialized, and the public must pay for it, while all profits are strictly private, threatening the savings and meager investments of working class people with a speculative bubble of historic proportions. If/when this bubble does burst, the techoligarchs will no doubt be bailed out with taxpayer money, while many ordinary people will be sent to the poorhouse (assuming they aren’t already there). While it might seem innocent enough, at this point, to use AI for assignments is to make an alliance with the forces of capital who explicitly and at every turn announce their disdain for humanity, democracy, the environment, and the lives of people like you and I. It is to choose subjection for temporary convenience. We must collectively build an alternative cultural and personal calculus. We don’t kneel.
Pardis Dabashi's AI Statement, Adopted from Marissa Fenley (BMC) and Olivia Stowell (UVA)
Al use in this class is, in every instance, unacceptable. Every assignment is designed in order to provide you with the opportunity to engage with the material and develop your own thinking without corrosive interference. Violations of this policy will result in a charge of plagiarism and treated accordingly (at the very least an automatic F on the assignment, without opportunity for resubmission).
These are some of the considerations behind this policy:
• This class is designed to improve your writing, thinking, and imaginative skills; if you're not writing, thinking, or imagining, you're not improving.
• Using Al opens up academic honesty issues, both because it raises the question of whether the work is "your" work, and also because tools like ChatGPT rely on taking uncredited material from scholars and writers (external source linked)
• Using Al does not produce reliably accurate results (external source linked x3)
• ChatGPT has serious negative environmental impacts. For instance, ChatGPT uses 7-10 times more energy than a Google search (external source linked), uses about one 16 oz water bottle for every 5-50 queries to cool its data centers (external source linked) and the C02 emissions for training a single large learning model are equal 125 round-trip flights between New York and Beijing (external source linked).
• Al is also putting massive strain on the energy grid (external source linked) and data centers are disproportionately built in poor and rural communities who are already the inheritors of environmental racism and classism (external source linked x3)
• OpenAl and other Al companies have exploited workers from the global south, especially workers in prisons and refugee camps, and workers in Kenya, the Philippines, Colombia, and elsewhere (external source linked x5)
• Al is actively destroying knowledge archives and historical materials in order to consume, and ultimately degrade, language and discourse (external source linked).
Most other English Professors here at Bryn Mawr have some version of this policy.