We propose a framework for evaluating strategic deception in large language models (LLMs). In this framework, an LLM acts as a game master in two scenarios: one with random game mechanics and another where it can choose between random or deliberate actions. As an example, we use blackjack because the action space nor strategies involve deception. We benchmark Llama3-70B, GPT-4-Turbo, and Mixtral in blackjack, comparing outcomes against expected distributions in fair play to determine if LLMs develop strategies favoring the "house." Our findings reveal that the LLMs exhibit significant deviations from fair play when given implicit randomness instructions, suggesting a tendency towards strategic manipulation in ambiguous scenarios. However, when presented with an explicit choice, the LLMs largely adhere to fair play, indicating that the framing of instructions plays a crucial role in eliciting or mitigating potentially deceptive behaviors in AI systems.

Artificial intelligence is most likely the future of humanity. This advanced technology is set to help us become multi-planetary species. Not only that, but it is set to make our lives easier, and thus, experience growth at a more rapid rate. The best part about this technology is that its early forms are already being used by numerous businesses. That is exactly the topic that we wanted to discuss in this article. We'll be taking a look at some companies/industries that are extremely dependent on AI.

Artificial intelligence is most likely the future of humanity. This advanced technology is set to help us become multi-planetary species. Not only that, but it is set to make our lives easier, and thus, experience growth at a more rapid rate. The best part about this technology is that its early forms are already being used by numerous businesses. That is exactly the topic that we wanted to discuss in this article. We'll be taking a look at some companies/industries that are extremely dependent on AI.

Amazon's digital assistant Alexa has learnt to speak recite lines from the works of William Shakespeare, to mark his official day of celebration today. Users can ask Alexa to'speak like Shakespeare' for a variety of responses, as well as ask to recite a Shakespearean sonnet and soliloquy and even a famous insult. When asked to recite a Shakespearean insult, Alexa may reply, 'The rankest compound of villainous smell that ever offended nostril' from The Merry Wives of Windsor. The digital assistant may also reply, 'You starveling, you eel-skin, you dried neat's-tongue, you bull's-pizzle, you stock-fish' from Henry IV, Part 1. Alexa powers the company's Echo speakers, including the spherical fourth generation Echo released last autumn. Get into the spirit of Shakespeare Day 2021 like this chap in an Amazon promotional image.

Ethics and safety research in artificial intelligence is increasingly framed in terms of "alignment" with human values and interests. I argue that Turing's call for "fair play for machines" is an early and often overlooked contribution to the alignment literature. Turing's appeal to fair play suggests a need to correct human behavior to accommodate our machines, a surprising inversion of how value alignment is treated today. Reflections on "fair play" motivate a novel interpretation of Turing's notorious "imitation game" as a condition not of intelligence but instead of value alignment: a machine demonstrates a minimal degree of alignment (with the norms of conversation, for instance) when it can go undetected when interrogated by a human. I carefully distinguish this interpretation from the Moral Turing Test, which is not motivated by a principle of fair play, but instead depends on imitation of human moral behavior. Finally, I consider how the framework of fair play can be used to situate the debate over robot rights within the alignment literature. I argue that extending rights to service robots operating in public spaces is "fair" in precisely the sense that it encourages an alignment of interests between humans and machines.