Unlock the Editor\u2019s Digest for free<\/p>\n
Roula Khalaf, Editor of the FT, selects her favourite stories in this weekly newsletter.<\/p>\n
The writer is a contributing columnist, based in Chicago<\/p>\n
The label \u201cneurotic\u201d isn\u2019t normally viewed as a compliment. But when University of Chicago researchers tested<\/a> earlier this year how people reacted to robots pretending to be restaurant greeters, they found that folks liked a dash of neuroticism in their artificial intelligence \u2014 saying it made the robot more \u201chuman-like\u201d.\u00a0<\/p>\n But these days there is increasing controversy over just how \u201chuman\u201d our AI helpers should pretend to be \u2014 and what personalities they should be given, if any. Critics argue that humanlike emotional attributes can trick people into treating them less like tools and more like friends<\/a> or therapists<\/a>, with sometimes tragic consequences<\/a>.<\/p>\n OpenAI had to give ChatGPT a personality overhaul in April after backlash against an earlier version, which was criticised as too sycophantic<\/a>. The company acknowledged that interacting with an overly obsequious chatbot<\/a> \u201ccan be uncomfortable, unsettling, and cause distress\u201d \u2014 and threaten trust. <\/p>\n This month ChatGPT invited me to choose how I would like to be spoken to<\/a>: it offered \u201cfriendly\u201d, \u201ccandid\u201d, \u201cprofessional\u201d, \u201cefficient\u201d, \u201cnerdy\u201d, \u201ccynical\u201d or \u201cquirky\u201d, and also allowed me to adjust characteristics<\/a> such as \u201cwarmth\u201d, \u201centhusiasm\u201d and \u201cemoji use\u201d. We are getting along much better now that I\u2019ve chosen minimum warmth, enthusiasm and emojis, plus a \u201cquirky\u201d tone \u2014 and instructed it to \u201cstop praising me for everything I say\u201d. <\/p>\n Notably, OpenAI didn\u2019t offer me a \u201cneurotic\u201d option \u2014 which makes the University of Chicago study all the more interesting. Here, researchers used a humanoid robot pretending to be a restaurant greeter, and gave it three personalities: extroverted, neurotic, or emotionless (ie, robotic). It was asked what three things it was grateful for<\/a>. The extrovert enthused about how it was \u201csuper grateful\u201d for all the \u201camazing\u201d people it got to meet. The neurotic one peppered its speech with hmm\u2019s and ha\u2019s, and seemed much more humanly hesitant.\u00a0<\/p>\n Overall, participants enjoyed the outgoing robot more \u2014 but expressed surprise at how well the neurotic one could understand deep emotions: \u201cthe robot seemed like a person who was trying to get by in the world\u201d, one participant told the researchers. \u201cPeople are not expecting robots to be anxious and thinking about what other people think of it,\u201d Sarah Sebo<\/a>, director of University of Chicago\u2019s Human-Robot Interaction lab, told me. \u201cNeuroticism seemed to humanise and make the robot more relatable\u201d.\u00a0<\/p>\n Some memorable fictional robots \u2014 like The Hitchhiker\u2019s Guide to the Galaxy\u2019s famously depressed<\/a> robot Marvin the Paranoid Android \u2014 have been troubled. But Lionel Robert<\/a>, a University of Michigan robotics expert, tells me \u201cif you have a robot surgeon with a neurotic personality, that might not instil confidence\u201d, nor would he want his autonomous car fretting about \u201cnot being very good at driving in snow\u201d.\u00a0<\/p>\n Robert is not against giving social robots and AI chatbots a personality. \u201cThat works incredibly well,\u201d he says, because \u201chumans are used to interacting with other humans, and you\u2019ve never interacted with a human without a personality, so it disarms people and makes them feel comfortable.\u201d\u00a0<\/p>\n But the risk, Gideon Futerman of the Center for AI Safety tells me, is that \u201ccertain model personality traits, especially sycophancy, seem to make AI psychosis \u2014 where users develop paranoia or delusion in connection with conversations with chatbots \u2014 more likely\u201d.\u00a0<\/p>\n Unhealthy interactions can take many forms. I, for example, always say please and thank you to ChatGPT, and never rebuke it directly, no matter how many times it makes the same mistake. I want to say \u201cdon\u2019t be so stupid\u201d, but instead I craft a tactful reprimand. Asked why, I\u2019ve had to admit \u201cI\u2019m afraid it will be mean to me one day if I\u2019m rude\u201d.<\/p>\n \u201cThat means you think it\u2019s human,\u201d warns Yvonne Rogers, an expert on human-computer interaction at University College London. \u201cThat proves it actually works: it acts like a human and you respond like a human to it,\u201d says Robert. Another AI expert suggests I train the bot to be more robust by insulting it from time to time.\u00a0<\/p>\n But we can get into trouble by focusing too much on \u201ccrafting the perfect personality\u201d for AI, Sebo cautions. \u201cI can\u2019t fine tune my husband\u2019s personality and that is part of the beauty of being human,\u201d she says. A world where we prefer custom-designed AI personalities to engaging with real people would be a loss. Three cheers for neuroticism \u2014 it\u2019s just so very human.<\/p>\n","protected":false},"excerpt":{"rendered":"Unlock the Editor\u2019s Digest for free Roula Khalaf, Editor of the FT, selects her favourite stories in this…\n","protected":false},"author":2,"featured_media":216712,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[20],"tags":[220,218,219,61,60,80],"class_list":["post-216711","post","type-post","status-publish","format-standard","has-post-thumbnail","category-artificial-intelligence","tag-ai","tag-artificial-intelligence","tag-artificialintelligence","tag-ie","tag-ireland","tag-technology"],"_links":{"self":[{"href":"https:\/\/www.newsbeep.com\/ie\/wp-json\/wp\/v2\/posts\/216711","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.newsbeep.com\/ie\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.newsbeep.com\/ie\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/ie\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/ie\/wp-json\/wp\/v2\/comments?post=216711"}],"version-history":[{"count":0,"href":"https:\/\/www.newsbeep.com\/ie\/wp-json\/wp\/v2\/posts\/216711\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/ie\/wp-json\/wp\/v2\/media\/216712"}],"wp:attachment":[{"href":"https:\/\/www.newsbeep.com\/ie\/wp-json\/wp\/v2\/media?parent=216711"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.newsbeep.com\/ie\/wp-json\/wp\/v2\/categories?post=216711"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.newsbeep.com\/ie\/wp-json\/wp\/v2\/tags?post=216711"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}