Dumb blonde no more: Barbie is getting smart with artificial intelligence
“Imagination, life is your creation. Make me walk, make me talk. …” Aqua lyrics have come true with Hello Barbie, a new wifi-connected doll that uses speech recognition technology to have real conversation with kids.
ToyTalk, a San Francisco startup, raised $31+ million to boost the toy market by creating The Kid’s Internet of Things. The company partnered with Mattel and launched a two-way communication Barbie in late November 2015.
Value creation and capture for kids: Pay premium to get a “real” friend
To be able to talk to Barbie has been the number one request of children for decades. Hello Barbie is connected to cloud via wifi; it records, remembers, processes and reacts to kids’ speech through artificial intelligence programming. Your “real friend” Barbie understands you, is excited to chat with you and can even help with tough life questions aka “What would you like to be when you grow up?”! Mattel and ToyTalk launched a dream Barbie just in time to capture part of the $6 billion Christmas toy market. Mattel positions Hello Barbie as a true premium product with a price point of $75 several times above the average Barbie price ($10 – $30).
Value creation and capture for ToyTalk: Data is king, Barbie serves as a pilot
ToyTalk creates value in two ways. First, whenever a kid engages with their product, ToyTalk’s kids speech recognition technology receives feedback and improves. Second, ToyTalk collects an immense database of unstructured data that could unlock deeper psychological understanding of what kids want, and how they react to products.
So far, ToyTalk has monetized its unique technology by selling proprietary iPad apps (the Winston Show, SpeakaZoo, SpeakaLegend), and now by licensing the technology to Mattel (Hello Barbie, Thomas & Friends Talk to You). ToyTalk is a software company, not interested in making physical products. They don’t have to be to capture value. Several toy majors (Mattel, Hasbro) have seen decreasing or flat sales over the last years of digitalization, and are ready to bet on The Kid Internet of Things.
ToyTalk will have to navigate carefully the delicate world of children data privacy
I see two other lucrative paths ToyTalk could pursue in the near future: commercialization of their data and products for parents. The data giants, Google and Amazon, are hungry for tapping into new customer interaction spheres. Google paid $3.2 billion for Nest, a learning thermostat technology and part of new smart-home systems. Will kids be their next focus? If ToyTalk will not be able to sell the children’s speech data or analysis per se, will the only option be to acquire the company?
Many parents have protested against the Hello Barbie model worried that their children’s data may be hacked or misused. Besides high security standards, ToyTalk could address this issue by creating value for parents themselves. The 21st century parents seek permanent control and surveillance. What if ToyTalk incorporated a streaming option that parents could use to see “if everything was fine with their little Mary ” while they are at work? What if ToyTalk database could serve as a kid’s audio diary?
ToyTalk has mastered the world of unstructured big data and found a very promising growth segment. It is clear that Hello Barbie is just a first swallow.
ToyTalk is a pioneer company of kids’ speech recognition and natural language processing. Kids’ speech is more complex that that of adults because of unexpected language usage and different voice frequencies. ToyTalk not only cracked the technology, it moved it a level up by creating a live improv for kids in cloud.
Smart toys are a great idea and have the potential to make lives easier for parents and more fun for kids. However, one big concern I have about the sector is uncertainty around the AI and speech recognition technology used in toys. Unlike with real family, friends and care takers, parents are not sure exactly what toys would say to their kids in different scenarios and what kind of positive/negative impact they would have on the kids. Unless the technology matures and offers consistently reliable responses, I think it would be hard to robots or smart toys to gain mass adoption quickly.
Thank you for your comment! I definitely agree that parents resistence may prevent a quick mass adoption. At the same time, Barbie & ToyTalk has done substantial amount of work in this area already. Barbie’s answer to “Do you believe in God?” (“I think a person’s beliefs are very personal to them.”) is very thoughtful way to teach kids about tolerance.
Check out this New York Times’ article: http://www.nytimes.com/2015/09/20/magazine/barbie-wants-to-get-to-know-your-child.html?_r=0
The surveillance angle has already been tackled- but needs to be mixed with good cyber security. There have been stories about hacked baby monitors with backdoors and such for years (see http://fusion.net/story/192189/internet-connected-baby-monitors-trivial-to-hack/ for a recent one). There are also questions about where the data from the barbie goes- does it go to ToyTalk’s cloud? Would the parents mind if they found out their children’s words were being relayed to Mattel? Or if they were used as spying products?
Granted, if Barbie is going to talk, she will be collecting data to better respond to “her” playmate. The toy companies are going to have to act before parents fear that Barbie is going to be dangerous and retreats to completely analog toys, or a video game system that does not receive voice commands without a peripheral such as the Playstation 4.
Thank you for your comment, Alexander!
Parents panicked already and created a petition on Campaign for Commercial-free Childhood portal earlier this year when the doll was presented (http://www.commercialfreechildhood.org/action/hell-no-barbie-8-reasons-leave-hello-barbie-shelf). Mattel and ToyTalk decided not to react and launched the product. There are restrictions on how the data can be used (research and development and data analysis); they cannot be used for marketing purposes at all (e.g., Barbie cannot say “Hey, Mary, buy Ken, I feel alone here.”). Plus parents have to agree with the terms and conditions before the product is connected to the ToyTalk’s cloud. However, I believe that there is a grey zone and that this zone will become regulated over time. I would expect a strict point of view of the regulators.
Not all parents have such an antagonistic point of view, obviously. If you go to YouTube, there are many public videos of little kids trying to communicate with Siri (such as https://www.youtube.com/watch?v=P8zG_gGmfmo). ToyTalk is more private in the sense that the access to the database is limited to the parents (they can even delete specific comments by their kid) and the data analysts…
What do you think about this?
The legal issue is always a problem. The “Terms of Service” question is an issue because some argue they are too long and legalistic to read, and some are hidden away (http://www.forbes.com/sites/oliverherzfeld/2013/01/22/are-website-terms-of-use-enforceable/). Because of this, even if a parent agrees to it there is a chance the parent will say “I did not see it in the terms of service.” The Supreme Court decided in a 5-4 decision that it is binding (when it comes to AT&T forcing people to use arbitration instead of litigation), but the 5-4 closeness (Scalia, Roberts, Kennedy, Thomas, and Alito in favor and all other opposed) suggests that the replacement of just one of the people who said “yes” with a “no” could make things more difficult for those who write EULAs.
The video was adorable. I understand and agree this is a big market segment, although I have never found a voice-activated device that I felt comfortable speaking to. Siri tends to drive me nuts and Amazon’s Echo feels heavily constrained. Kids may love it, and there’s money in it. I still have non-security questions however. How varied is Barbie’s dialogue and how long will it take before kids find Barbie’s limit? It looks likethere’s quite a bit now, but how long? Will price be an issue? At $75 she seems a tad expensive (and 2x-6x the price of a “normal” Barbie doll), although I have never bought a toy for young girl so I do not know if people normally buy single Barbies or not. She is certainly a step in an interesting direction, but between stories about baby monitors’ lack of security and everything else I see her more as a step for other people to beta test than something to invest in right now. Should be interesting to watch.
Great points both on the legal side and the business side.
Regarding the price point, it seems to be within a range of the most wanted toys for Christmas 2014 (http://www.techtimes.com/articles/15707/20140916/christmas-shopping-2014-target-lists-30-toys-kids-want.htm), more on the upper side for kids below the age of 10. It would definitely be a premium toy with a limited reach for this Christmas, though this may be exactly what the authors wanted to do. The core fans will serve to improve the software (and may not mind some initial issues). The mass adoption will happen once the software is fixed (and the price point can be lowered).
To me, ToyTalk’s approach and business model is further evidence that “software is eating the world.” Their capabilities in natural language processing, machine learning and ability to make once old toys new and exciting for children again demonstrates the unique value software can add to undifferentiated hardware products. Just as apps and digital personal assistants differentiate smartphones today, so ToyTalk’s software can be the next phase of innovation and value creation in the toy industry. I would predict that, if not acquired by a toy manufacturer or other tech company, ToyTalk will be able to capture more and more value from the toy brands in the future.
Thank you for your comment! I also immediately thought about the article “Software is eating the world”. I believe that as we have difficulties imagining our phones without interactive apps these days, our kids will one day not understand a value proposition of a “non-smart” Barbie.
I would recommend ToyTalk to keep the company ownership for the next 1-2 years as the expected value of their business is set to reach an exponential growth.