机器人Jibo

    At first glance, Jibo looks a bit like Wall-E’s robot girlfriend. Both Jibo, a real robot, and Wall-E’s girlfriend, the fictional Pixar character, have the look of a futuristic Apple product: reflective white plastic, round curves, a black screen for a “face,” and smooth swiveling movements.

    But Jibo’s raison d’être is slightly more in line with Rosie, the robot maid from the 1960s animated television series The Jetsons, and its operating system is more akin to the one employed by Samantha, the artificially intelligent character from the 2013 Spike Jonze film Her. (One key difference: Jibo is male, according to its makers.)

    Jibo is described as a “family robot” because it is able to see, hear, speak, learn, and help families with a variety of tasks around the house. It—he?—can “relate” by expressing itself in natural language, using “social and emotive cues so you understand each other better.” Jibo is meant to be a companion.

    It’s the creation of a team of robot architects, cloud computing engineers, animators, conversational technologists, and human-robot interaction engineers. Jibo, Inc. is backed by $5.59 million in venture funding from investors including Charles River Ventures, Fairhaven Capital Partners, Osage University Partners, and angel backers.

    So far, people like Jibo. A crowdfunding campaign, launched last month, raked in more than $1.5 million from more than 3,500 people, handily surpassing its $100,000 target. (The Boston-based company does not expect to ship its first units, priced at $499 each, until the 2015 winter holiday season. The crowdfunding campaign is designed to get developers excited about building apps for the robot, it said.)

    Naturally, I had to meet Jibo. Off to a hotel room in Midtown Manhattan, then, where two Jibos and Dr. Cynthia Breazeal, the robot’s creator, awaited me. The robot is not yet fully functioning, it turns out. I watched a prepared demo where Jibo, about a foot tall, turned to look me in the eye. This was disarming at first, as if I was being followed by a security camera. Once he started talking to me, it began to feel more natural—as natural as a robot in a 1980s science fiction movie, anyway. Unlike his lesser robotic peers, or, say, a smartphone, Jibo did not rudely buzz or ding when there was a new message to communicate to me. He politely said, “Excuse me, Erin,” and waited for me to respond before continuing.

    In the room, Jibo showed off his swiveling, spinning and leaning moves to me, along with some of the programs he’ll feature. He ended his performance with a cheesy joke, and his eyes turned to tiny half-moons when he laughed at the punch line.

    乍一看，Jibo长得有点像皮克斯动画工作室(Pixar)创作的《机器人总动员》(Wall-E)里那个主人公的女友Eve。Jibo和Eve的外观都颇具未来主义范儿，看上去像是苹果公司(Apple)的产品：光滑的白色塑料外壳，圆润的曲线，一张充当“脸”的黑色屏幕，转动灵活。

    不过从本质上看，Jibo其实更像上世纪60年代动画片《摩登家庭》（The Jetsons）里的机器人女佣Rosie。它的操作系统更类似于2013年斯派克•琼斯的电影《她》(Her)中的虚拟人工智能角色Samantha使用的那种。（但根据其制作者介绍，一个关键的区别是，Jibo其实是个男孩。）

    Jibo被称做一款“家庭机器人”，因为它具有看、听、说、学等功能，而且可以帮助我们干许多家务活。它（或者说“他”？）可以使用自然语言来表达自己，并且可以“使用社交性和感性的暗示，让你们更好地理解对方。”总之，Jibo致力于成为一个居家伙伴。

    它是一群机器人设计师、云计算工程师、动画工程师、会话技术专家和人机互动工程师的心血之作。Jibo公司也从查尔斯河风险投资公司（Charles River Ventures）、菲尔海文资本合作公司（Fairhaven Capital Partners）、奥塞治大学合伙公司（Osage University Partners）和天使投资人那里获得了559万美元的融资。

    到目前为止，人们都很喜欢Jibo。上个月Jibo推出了一个众筹项目，很快就从3500多人那里筹集到了150多万美元，轻而易举地超过了该公司预设的10万美元的目标。（这家总部位于波士顿的公司预计，第一批定价为499美元的产品要等到2015年冬的假日季才能上市。该公司表示，推出这个众筹项目是为了激发软件开发者为Jibo设计应用程序的兴趣。）

    当然，我也要去见一下Jibo。在曼哈顿市中心的一家酒店的房间里，两个Jibo机器人和它们的发明者辛西娅•布雷西亚博士正在等着我。事实证明，Jibo的功能还没有充分完善。就在我观看一段准备好的演示视频时，大概一英尺高的Jibo突然扭过头来直勾勾地看着我。这种眼神一开始让我觉得有点紧张，好像是被一台安保摄像机盯着。一旦开始说话，它给人的感觉就自然多了——至少像80年代科幻片里的机器人一样自然。和那些自动化程度不高的智能设备相比（比如智能手机），准备向我传递新信息时，Jibo不会粗鲁地“嗡”或“叮”一声，而是会礼貌地说道：“打扰了，艾林”，然后等我做出反应后，才会继续说话。

    在房间里，Jibo向我展示旋转，偏头等动作，以及一些他搭载的一些程序。最后他以一个很劲爆的笑话结束了表演。当这个笑点惹得他大笑的时候，Jibo眼睛眯成了一个小小的月芽。

    Jibo can perform a number of functions. He can tell children’s stories and snap family photos using face recognition. He can place Skype calls and handle communications for which you would normally use a phone. Jibo is meant to stay in the home, perched on a table or countertop, and a demo video shows him greeting a single man when he comes home from work and offering to order Chinese takeout. In another scene, Jibo is hanging out while a woman kneads bread. He chimes in to remind her that her daughter is picking her up soon. “Thanks, Jibo,” the woman responds, not unlike Jane Jetson talking to Rosie.

    Jibo can be considered the next logical step past today’s “telepresence” robots, which work only by connecting a smartphone or tablet—a brain, if you will—to a mobile base. For example, Romo augments your cell phone with rubber tank treads, though it requires a tablet or another phone to serve as a remote controller. Ubooly is a plush children’s toy in which parents can insert their cell phone for playtime. The Double telepresence robot, essentially an iPad on top of a Segway, allows people to feel physically present in meetings and move around the office when they’re working remotely. It’s a bit like Max Headroom on a broomstick and, to be frank, a little silly in practice.

    Jibo works with smartphones, but Breazeal chose to give the robot its own brain, rather than rely on a smartphone. The smartphone would have limited the robot’s capabilities, she says. As it turns out, people don’t like to put their phones into a robot anyway. They prefer to keep it on hand, Breazeal says.

    Whether that can make a difference—or translate to sales of in-home robots—is up for debate, but if anyone can figure this out, it’s Jibo’s inventor. Breazeal has dedicated her career to social robots, starting as a grad student at M.I.T. When she was younger, she didn’t understand why NASA was sending robots to Mars but they still hadn’t arrived in people’s homes. It’s because those robots weren’t designed to be social, she reasoned. Breazeal went on to build the first a social robot, which was called Kismet and intended for children. She has since published numerous studies on social robotics and in 2010 delivered a TED talk on the subject. People respond to human-like robots the same way they respond to people, she argued, and robots with the ability to convey expression increase empathy, engagement, and collaboration among people in a way that a robot with a flat demeanor cannot.

    An estimated 3 million service robots, which are intended for personal and domestic use, were sold in 2012, according to the International Federation of Robotics, representing sales of $1.2 billion. The IFR predicts 22 million robots to be sold through 2016.

    Jibo is purposely designed to not resemble a human, Breazeal says. The goal is to create what she calls a humanized experience, “because that’s what empowers people,” she says. Robots that try to look like human beings end up being a little too science fiction.