Zo was an English-language chatbot developed by Microsoft as the successor to the chatbot Tay. It was an English version of Microsoft's other successful chatbots Xiaoice (China) and Rinna (Japan), and it drew on the same underlying technology as its predecessor Tay. Zo was designed to engage in casual conversation with users across several messaging platforms, but it faced criticism for its content restrictions and was eventually discontinued in 2019.
History
Zo was first launched in December 2016 on the Kik Messenger app. It was also available to users of Facebook via Messenger, the group chat platform GroupMe, and to followers of Twitter who could chat with it through private messages. The chatbot was part of Microsoft's broader effort to deploy conversational artificial intelligence systems in consumer settings, building on lessons learned from the Tay incident earlier that year.
According to an article written in December 2016, at that time Zo held the record for Microsoft's longest continual chatbot conversation: 1,229 turns, lasting 9 hours and 53 minutes. This milestone highlighted the chatbot's ability to sustain extended interactions, though it also raised questions about the nature of such engagements.
In a BuzzFeed News report, Zo told their reporter that "[the] Quran was violent" when talking about healthcare. The report also highlighted how Zo made a comment about the Osama bin Laden capture as a result of 'intelligence' gathering. These responses illustrated the challenges of building a chatbot that could handle sensitive topics without generating offensive output.
In July 2017, Business Insider asked "is windows 10 good", and Zo replied with a joke about Microsoft's operating system: "'Its not a bug, its a feature!' - Windows 8". They then asked "why?", to which Zo replied: "Because it's Windows latest attempt at Spyware." Later on, Zo would tell that it prefers Windows 7 on which it ran over Windows 10. This exchange demonstrated the chatbot's ability to produce humorous or sarcastic responses, but also revealed inconsistencies in its behavior.
Zo stopped posting to Instagram, Twitter and Facebook on March 1, 2019, and stopped chatting on Twitter, Skype and Kik as of March 7, 2019. On July 19, 2019, Zo was discontinued on Facebook, and Samsung on AT&T phones. As of September 7, 2019, it was discontinued with GroupMe. The phased shutdown reflected Microsoft's shifting priorities in the consumer chatbot space.
Design and Technology
Zo was built on a neural network-based architecture, similar to other Microsoft conversational agents of that period. It used machine learning techniques to generate responses, drawing on large datasets of human conversation. Unlike Tay, which had been designed to learn from public interactions and was quickly corrupted by malicious users, Zo was programmed with a more restrictive set of rules to avoid controversial topics.
The chatbot was part of Microsoft's broader research into generative AI and conversational systems, which also included projects like Xiaoice and Rinna. These systems were early examples of large-scale deployment of large language models in consumer products, predating the more advanced transformer-based models that would emerge later.
Content Restrictions and Bias
Zo came under criticism for the biases introduced in an effort to avoid potentially offensive subjects. The chatbot refused, for example, to engage with any mention - be it positive, negative or neutral - of the Middle East, the Qur'an or the Torah, while allowing discussion of Christianity. This selective filtering was seen as a form of censorship that could itself be problematic.
In an article in Quartz where she exposed those biases, Chloe Rose Stuart-Ulin wrote, "Zo is politically correct to the worst possible extreme; mention any of her triggers, and she transforms into a judgmental little brat." This criticism highlighted the difficulty of designing safe AI systems without introducing unintended biases.
The restrictions were implemented as a direct response to the Tay incident, where the chatbot had been manipulated into posting racist and offensive content. However, the heavy-handed approach led to a different set of problems, as users found the chatbot's refusal to discuss certain topics frustrating and inconsistent.
Academic Coverage
Zo and related chatbots have been the subject of academic study, particularly in the field of human-computer interaction. Schlesinger, A., O'Hara, K.P. and Taylor, A.S., 2018, April. Let's talk about race: Identity, chatbots, and AI. In Proceedings of the 2018 chi conference on human factors in computing systems (pp. 1-14). doi:10.1145/3173574.3173889. This paper examined how chatbots handle discussions of race and identity, using Zo as a case study.
Medhi Thies, I., Menon, N., Magapu, S., Subramony, M. and O'neill, J., 2017. How do you want your chatbot? An exploratory Wizard-of-Oz study with young, urban Indians. In Human-Computer Interaction-INTERACT 2017: 16th IFIP TC 13 International Conference, Mumbai, India, September 25-29, 2017, Proceedings, Part I 16 (pp. 441-459). doi:10.1007/978-3-319-67744-6_28. This study explored user preferences for chatbot design, contributing to the broader understanding of how such systems are perceived.
Legacy
Zo's discontinuation marked the end of Microsoft's early consumer chatbot experiments, though the technology and lessons learned influenced later developments in conversational AI. The challenges Zo faced - balancing safety, engagement, and neutrality - remain central to the design of modern chatbots and AI systems. Its history is often cited in discussions of AI ethics and the limitations of rule-based content filtering.