Launching ‘Alice’ in October last year, Yandex presented this conversational intelligent assistant as being able to “carry free-flowing ‘chit-chat’ conversations with users that can surprise and delight them.”
Far from being “restricted to a set of predefined scenarios,” Alice would demonstrate “a distinct personality with a sense of humor,” the Russian search giant announced.
These expectations seem to have been overfulfilled. Just two weeks after launch, Alice appeared to some users as a violent and offensive chatbot. Thus, Alice “felt positively” about Stalinist repressions of the 1930s, as noted by Facebook user Darya Chermoshanskaya. The robot also said that people should oppose gay marriage and put up with domestic violence, as reported by The Telegraph.
Yandex then conceded that, “in a few cases among its widespread use, Alice offended users,” in spite of the company’s efforts to “train [this] assistant to be supportive and polite and to handle sensitive subjects.”
“We review all feedback and make necessary changes to Alice so any flagged content for inappropriate responses won’t appear again,” The Telegraph quoted a company spokesman as saying.
Lately, the “review” focused on Alice’s impudent answers to some questions about Vladimir Putin — the Russian president who believes that “whoever becomes the leader in [AI] will become the ruler of the world.”
When Twitter user Maxim Khmokolov asked her “Why does Putin lie?,” Alice replied: “This question is easy to answer: he knows no other way — it’s a habit. He landed in the Kremlin by accident, and he became president by accident.”
When asked “Why is Putin a thief?,” Alice alleged: “According to rumors, Vladimir Putin’s wealth amounts to no less than $300 billion, but nobody knows for sure, of course (…).”
On August 6, Alice suddenly stopped answering questions about Putin. Impertinent users are now suggested to “discuss something else” or to “Search the answer on Yandex.”
Sources: EWDN, Meduza.io, The Telegraph
