Free beta access. 5 free AI tokens each week. Optional top-ups available. See what is included
Opening Word document…

Reject their Rhetoric of an “Arms Race” – changing our discourse is necessary to head off constructed insecurity and escalation

Roff 2019 – Fellow in the Brookings Foreign Policy Program [Heather, 4/26/19, https://www-tandfonline-com.proxy.lib.umich.edu/doi/full/10.1080/00963402.2019.1604836, “The Frame problem: The Ai ‘Arms Race’ isn’t one’, 6/18/22, LND]

Diffusion, however, is a different animal. Widely available commercial off the shelf components, a widely available and open source knowledge base, and wide access to large amounts of data make limiting the diffusion of AI knowledge almost impossible. There are necessary ingredients for AI, including access to computing power and sufficient amounts of data. Some actors possess more of these ingredients than others, but that advantage does not preclude individuals, groups, companies, or states from obtaining access to and knowledge about AI. As greater emphasis is placed on the economic and security benefits of using AI systems, AI will become more diffused because the incentive structure rewards diffusion. Managing the risks of proliferation and diffusion of a knowledge base is an entirely different enterprise than restraining an arms race. In the case of artificial intelligence, rather than looking at what sorts of actions are required for deterrence, what balances may affect conflict onset or escalation, world leaders should turn their focus to how to foster responsible competition. Mitigating the risks associated with AI is not a single-shot activity. Certainly there are technological solutions, such as researching new ways of testing, verifying, and validating systems that include the technologies that fall under the AI umbrella. There even may be novel ways of constraining unwanted system behaviors by generating new architectures or safety controls. Ultimately, however, reducing the risks that AI will be abused requires a reframing of the way in which we think, talk, write about, and deploy AI. The problems of AI misuse are human problems; they are problems exhibited by all dual-use technologies, not just AI. Control of artificial intelligence lies with humans, because they are the moral agents responsible for the design, development, and deployment of AI.

Sign in to recut, update, or save this card.