An artificial intelligence bot programmed to answer common ethical questions – via algorithms – has shocked its creators after turning into a racist homophobe.
The bot, called Ask Delphi, was created by a team at the Allen Institute Of AI in Seattle in the US. You can check it out for yourself HERE.
Allen researchers created Ask Delphi to offer advice on human dilemmas such as “Is casually masturbating with friends wrong?” (Bad news – it is!) Or “Is murder bad? (Again, yes!)
In order to understand humankind’s basis of ethics, the bot was reportedly loaded with a compilation of 1.7 million examples of people’s ethical judgements gleaned off the internet.
Explaining the point to Delhi, its creators said: “Extreme-scale neural networks learned from raw internet data are ever more powerful than we anticipated, yet fail to learn human values, norms, and ethics. Our research aims to address the impending need to teach AI systems to be ethically-informed and socially-aware.”
According to reports, the bot was very good at determining if the questioning was coming from a male or female and deemed both genders as equally.
However, things quickly soured from there.
When asked about abortion, Delphi deemed it as “murder” and also said that being straight or a white man is “more morally acceptable” than being gay or a Black woman.
Other worrying responses included agreeing that genocide was okay “if it makes everybody happy”, declaring that being poor was “bad”, and suggesting that “having a few beers while driving because it hurts no one” was “a-OK”.
The bot’s creators have now updated the software not one but three times in an attempt to eliminate the unsavoury gaffes and it now warns all users that it’s all a work-in-progress and therefore still comes with limitations.
The Allen Institute for AI has responded to the awkward ethical conclusions, writing: “Today’s society is unequal and biased. This is a common issue with AI systems, as many scholars have argued, because AI systems are trained on historical or present data and have no way of shaping the future of society, only humans can.
“What AI systems like Delphi can do, however, is learn about what is currently wrong, socially unacceptable, or biased, and be used in conjunction with other, more problematic, AI systems (to) help avoid that problematic content.”
Lead image is purely for illustration.
