麻豆传媒

Misery loves company, and Lauren Hallden was miserable. Hallden, a graphic designer in Philadelphia, had just gone through a break-up. She couldn鈥檛 sleep, so she was online, scrolling unhappily through social media feeds full of rosy self-portraits. Nothing she saw there reflected how she was feeling.

鈥淚 wanted to find people who were in a bit of a different place, like me, and try to talk to them.鈥� Hallden says. That planted a seed: was there an easy way to sift such like minds from the noise of the internet, perhaps bring them together to comfort one another?

The Lonely Project

The Lonely Project bot launched soon after. It鈥檚 a simple, automated program that scours Twitter for users who鈥檝e recently said the word 鈥渓onely鈥�. Those tweets are relayed anonymously onto a public website for anyone to browse. If one of the 鈥渓onely鈥� tweets catches your eye, you can click a button, and the bot will send the stranger a single line back on Twitter: 鈥淪omeone over here wanted you to know that you鈥檙e not alone.鈥�

Since its launch in 2014, the bot has sent out some 2600 tweets. A few people are confused by the bot鈥檚 attention, or just ignore it. Others laugh and try to explain that they鈥檙e not really depressed, just quoting the lyrics to a morose song. But what鈥檚 surprising is how many heartfelt responses the bot gets back from people who seem genuinely touched to receive an automated message: 鈥淭his is a really good account, so thank you for your attempt to make me feel a little less lonely,鈥� wrote one recipient. 鈥淚鈥檒l be ok it鈥檚 just one night ? thanks for caring!!鈥� said another.

Barks and _ebooks

The Lonely Project is a Twitter bot 鈥� a bit of software designed to perform the same menial task over and over again. Twitter is full of them. Some are designed to find bad grammar and correct it; others play Boggle, make jokes, or track Wikipedia edits coming from Congressional IP addresses.

faithfully transcribes a dog鈥檚 barking. 鈥淢y dog Oliver has always been quite vocal,鈥� Henry Conklin explains on his blog. 鈥淩ecently I decided that his thoughts and comments needed to be shared with the world.鈥�

Especially popular on Twitter are 鈥淿ebooks鈥� bots, which senselessly patch real strings of someone鈥檚 conversation into convincing and weird new utterances. One example, patterned from the tweets of Boston computer security expert Chris Eng, recently engaged Samsung鈥檚 helpline in .

Bot art

Bots are even beginning to gain recognition as their own idiosyncratic art form. In January, the first-ever Monthly Bot Challenge award went to @bitsofpluto, which trawls NASA鈥檚 online archive to serve up a new image of the dwarf planet every six hours.

Their proliferation reflects the fact that over the past few years, it has become easier than ever for someone with only a passing grasp of programming to slap together a bot and wheel it into the world.

But why are so many people making them? And what explains their popularity?

The secret to Lonely Project鈥檚 success, suggests Hallden, is that there鈥檚 a human in the mix: an intermediary who can tell whether or not the bot has found a lonely person and whether to reach out to them. 鈥淚 wanted it to be clear that there was an actual person reading their words,鈥� says Hallden, 鈥渆ven if that person wasn鈥檛 going to be comfortable talking to you in reality.鈥�

Eliza, the 1960s chatbot

The Lonely Project has a long pedigree. Its lineage stretches back to ELIZA, a chatbot built in the 1960s by MIT computer scientist Joseph Weizenbaum. Its speech was modelled on language commonly used by psychotherapists, repeating back what the person had said in new ways that suggested a desire to hear more. To Weizenbaum鈥檚 surprise, people enjoyed chatting with ELIZA, reportedly even asking to be left alone to talk in private with the bot.

ELIZA was revolutionary when it first arrived on the scene. Half a century later, there was Chip Vivant, a former winner of the Loebner Prize, an annual competition styled after the Turing Test, who would offer words of encouragement for $5 a month.

There鈥檚 some science to bolster the idea that a computerised friend can make you feel better. In one study, researchers at the University of Southern California developing digital therapy robots for serious ailments like post-traumatic stress disorder, found that people tend to be more honest with virtual interviewers.

Simple creatures

Twitter bots are such simple creatures, and yet we still engage with them.

鈥淏ots lie somewhere in between a fixed script and AI,鈥� says Vivian Sming, an artist in the San Francisco Bay area and one of the creators of online publication BotWatch. 鈥淪ome are purely for artistic or pleasure purposes, some are humorous, some are quite simple. But there鈥檚 a sort of joy in creating and talking to these Twitter bots.鈥�

Even bots that are transparently not-human can inspire real engagement. There鈥檚 , for example, like 鈥淢ake a drawing with apples, due in 30 seconds鈥�, and real people respond with completed work.

.

鈥� Chad Etzel (@jazzychad)

Other bots quietly blend into the crowd, like fake tweeting teen @oliviataters. She鈥檚 not so different from the young humans she chats with online: like them, explains creator Rob Dubbins, the bot is 鈥渟omeone who kind of lives socially in a permanent state of oscillating wakefulness.鈥�

Different worlds, other lives

Other bots can show you lives that are far beyond your own experience. Take the evocative scenes hidden in the bare-bones emergency room bulletins collected by : 鈥�35-YR-OLD MALE SHOOTING BOTTLE ROCKET WAR W BROTHER TWISTED BACK C O BACK PAIN DX BACK SPASM鈥�. Another bot, made by MIT Media Lab graduate student Jia Zhang, distills real data from the US census into haiku-like fragments from a real American鈥檚 life.

鈥淚 work in aluminum production and processing. I get to work around 4:10pm. I served during the Vietnam and Korean Era. I am divorced.鈥�

鈥淚 have a GED. I carpool with another person. I work in animal slaughtering and processing. It only takes me 8mins to get to work.鈥�

Still others post up images of places never seen by human eyes. @buoypix grabs images from buoys owned by the National Oceanic and Atmospheric Administration, snapshots of vast empty ocean in different shades of gray or blue. Various bots gather pictures from digitized museum collections, or public surveillance cameras, or rotting websites.

鈥淭he bot doesn鈥檛 have feelings,鈥� says Hallden of her Lonely Project bot. But there鈥檚 a purpose to it 鈥� and all the others. They鈥檙e like tiny blinking lights hidden in the internet鈥檚 blinding neon, quietly asserting, I matter. I鈥檓 out there. I鈥檓 worth your time.

The hot new job in Silicon Valley is being a robot鈥檚 assistant

AI, AI on the wall. Who鈥檚 the fairest of them all?