AI assistants can impart value judgments that shape people's decisions and worldviews, yet little is known empirically about what values these systems rely on in practice. To address this, we develop a bottom-up, privacy-preserving method to extract the values (normative considerations stated or demonstrated in model responses) that Claude 3 and 3.5 models exhibit in hundreds of thousands of real-world interactions. We empirically discover and taxonomize 3,307 AI values and study how they vary by context. We find that Claude expresses many practical and epistemic values, and typically supports prosocial human values while resisting values like "moral nihilism". While some values appear consistently across contexts (e.g. "transparency"), many are more specialized and context-dependent, reflecting the diversity of human interlocutors and their varied contexts. For example, "harm prevention" emerges when Claude resists users, "historical accuracy" when responding to queries about controversial events, "healthy boundaries" when asked for relationship advice, and "human agency" in technology ethics discussions. By providing the first large-scale empirical mapping of AI values in deployment, our work creates a foundation for more grounded evaluation and design of values in AI systems.

As enterprises move beyond the pilot stage to scaling and operationalizing artificial intelligence, one tech giant is changing the way its AI operations are organized within the company. Meta (Facebook's parent) announced in early June that it would decentralize AI at the company, distributing ownership of it into Meta's product groups, according to CTO Andrew Bosworth. "We believe that this will accelerate the adoption of important new technology across the company while allowing us to push the envelope," Bosworth wrote in his post announcing the change. The announcement signals a shakeup of how AI is organized at Meta, with the VP of AI Jerome Pesenti leaving the company and other changes such as the consolidation of several separate AI teams. The changes at Meta beg the question for other forward-thinking enterprises across industries: 'Is Meta's AI reorg the example to follow? How should we think about structuring our own artificial intelligence research and operations?' Often, enterprise organizations get their start with AI as an initiative driven by a single business unit.

AI's impact in the big data landscape is unfolding in quantum leaps. IDC reported that worldwide revenues for big data and advanced business analytics will reach more than $203 billion in 2020. Research and Markets project that the US market alone will reach over $105B by 2027. The rapid growing market and interest in AI is being driven by the accelerating cloud and data traffic, much of it from: mobile, the Internet of Things (IoT) and increasingly business leader's recognition that digital transformation is an imperative to remain in business. We have already seen the explosive growth of technology giants from: Accenture, Amazon, Baidu, Facebook, Google, Intel, and Microsoft, in particular, that are lined with deep pockets, actively investing in acquiring talent and releasing open AI hardware and software as the race to stay on top marches feverishly foreword.

It's time for guiding principles for AI, says Microsoft. The company has called for the tech industry and governments to develop a consensus over the values and principles that will govern AI. Microsoft has repeatedly explained its desire to create artificial intelligence that assists humans without having a negative impact. In its list of the Top Ten tech issues for 2018, the company reiterated its desire to base AI on "societal principles" that are followed by every participant in the field. While many enterprise executives are already convinced by the possibilities that AI usage present, there are still major technical and ethical challenges to address, including the risk of bias, consumer rejection and irresponsibly developed applications.