Despite all efforts during almost half a century, most chatbots are still easily uncovered, but over the next decades they will definitely get smarter and finally we will distinguish human beings by them giving us silly answers as opposed to the much smarter chatbots. All of this will really start accelerating as soon as one single chatbot is smarter than one single human being. They will then be able to learn from each other, instead of learning from human beings, their knowledge will explode and they will be able to design even better learning mechanisms. In the long run, we will learn language from chatbots instead of the other way around.
This is where most applications of NLP struggle, and not just chatbots. Any system or application that relies upon a machine’s ability to parse human speech is likely to struggle with the complexities inherent in elements of speech such as metaphors and similes. Despite these considerable limitations, chatbots are becoming increasingly sophisticated, responsive, and more “natural.”

A rapidly growing, benign, form of internet bot is the chatbot. From 2016, when Facebook Messenger allowed developers to place chatbots on their platform there has been an exponential growth of their use on that forum alone. 30,000 bots were created for Messenger in the first six months, rising to 100,000 by September 2017.[8] Avi Ben Ezra, CTO of SnatchBot, told Forbes that evidence from the use of their chatbot building platform pointed to a near future saving of millions of hours of human labour as 'live chat' on websites was replaced with bots.[9]

In 1950, Alan Turing's famous article "Computing Machinery and Intelligence" was published,[7] which proposed what is now called the Turing test as a criterion of intelligence. This criterion depends on the ability of a computer program to impersonate a human in a real-time written conversation with a human judge, sufficiently well that the judge is unable to distinguish reliably—on the basis of the conversational content alone—between the program and a real human. The notoriety of Turing's proposed test stimulated great interest in Joseph Weizenbaum's program ELIZA, published in 1966, which seemed to be able to fool users into believing that they were conversing with a real human. However Weizenbaum himself did not claim that ELIZA was genuinely intelligent, and the introduction to his paper presented it more as a debunking exercise:

According to the Journal of Medical Internet Research, "Chatbots are [...] increasingly used in particular for mental health applications, prevention and behavior change applications (such as smoking cessation or physical activity interventions).".[48] They have been shown to serve as a cost-effective and accessible therapeutic agents for indications such as depression and anxiety.[49] A conversational agent called Woebot has been shown to significantly reduce depression in young adults.[50]
Bots are also used to buy up good seats for concerts, particularly by ticket brokers who resell the tickets.[12] Bots are employed against entertainment event-ticketing sites. The bots are used by ticket brokers to unfairly obtain the best seats for themselves while depriving the general public of also having a chance to obtain the good seats. The bot runs through the purchase process and obtains better seats by pulling as many seats back as it can.
Although Weizenbaum created his ELIZA thirty years before Internet became familiar to the general public, his creation is still alive and accessible to everyone. Watch the following video created by a youtube user IanProCastsCoUk, and see how the javascript version of Eliza emulates a Rogerian psychotherapist, responds on questions and leads simple conversations.

Pop-culture references to Skynet and a forthcoming “war against the machines” are perhaps a little too common in articles about AI (including this one and Larry’s post about Google’s RankBrain tech), but they do raise somewhat uncomfortable questions about the unexpected side of developing increasingly sophisticated AI constructs – including seemingly harmless chatbots.
The first formal instantiation of a Turing Test for machine intelligence is a Loebner Prize and has been organized since 1991. In a typical setup, there are three areas: the computer area with typically 3-5 computers, each running a stand-alone version (i.e. not connected with the internet) of the participating chatbot, an area for the human judges, typically four persons, and another area for the ‘confederates’, typically 3-5 voluntary humans, dependent on the number of chatbot participants. The human judges, working on their own terminal separated from one another, engage in a conversation with a human or a computer through the terminal, not knowing whether they are connected to a computer or a human. Then, they simply start to interact. The organizing committee requires that conversations are restricted to a single topic. The task for the human judges is to recognize chatbot responses and distinguish them from conversations with humans. If the judges cannot reliably distinguish the chatbot from the human, the chatbot is said to have passed the test.
Interface designers have come to appreciate that humans' readiness to interpret computer output as genuinely conversational—even when it is actually based on rather simple pattern-matching—can be exploited for useful purposes. Most people prefer to engage with programs that are human-like, and this gives chatbot-style techniques a potentially useful role in interactive systems that need to elicit information from users, as long as that information is relatively straightforward and falls into predictable categories. Thus, for example, online help systems can usefully employ chatbot techniques to identify the area of help that users require, potentially providing a "friendlier" interface than a more formal search or menu system. This sort of usage holds the prospect of moving chatbot technology from Weizenbaum's "shelf ... reserved for curios" to that marked "genuinely useful computational methods".