Microsoft is launching its AI chatbot Bing onto mobile devices globally Wednesday, offering users the chance to experience the attention-grabbing software that only last week sounded alarms following several wild outbursts.

Bing is now available on Microsoft's Edge mobile browser and dedicated Bing app on iOS and Android devices for those who have been accepted from the waitlist. Microsoft says more than one million people in 169 countries have previewed the AI-powered chatbot since its debut two weeks ago.

Mobile users on the Bing app will be able to start a live chat session with the bot with the touch of a finger, allowing them to ask a multitude of questions through text or by voice.

Answers may come in three varying formats, including bullet points, short-form responses, or in a traditional text format. This feature was added along with shortened conversation lengths, a response to the feedback that users felt the bot insulting, threatening, or bullying them when the conversation reached a certain length.

Microsoft is also pushing its bot onto its Skype video conferencing and messaging platform, allowing users to add the AI to group chats and ask it questions. In its release, Microsoft gave the example of chatting with a family member about an upcoming vacation and using the chatbot to provide recommendations on where to go, eat, etc.

Built by Microsoft partner OpenAI, the Bing chatbot experienced some notable turbulence upon its debut on Feb. 7. Microsoft has been criticized by many for rushing the bot's launch, and critics point to claims the AI answered questions incorrectly, issued threats, and told of desires to steal nuclear codes or to be alive, as evidence to this belief.

Microsoft's AI system is part of a wave of new technologies that have mastered human language and grammar after being programmed with a massive database of books and online writings.

These systems can develop recipes and compose songs and emails on command, or concisely summarize concepts with information found across the internet. But despite the increased prevalence of these chatbots, and stirring beliefs that they are nearing human-like sentience, they remain error-prone and cumbersome.