Young Woman Using AI Device

How AI Will Shape Life in the Home

Thanks to the rapid advancement of technology and a healthy global economy in which companies compete to develop the most impressive and compelling consumer products, the future of life in the home is shaping up to be characterized by artificial intelligence. Google, for instance, has shifted its business to focus on so-called “ambient computing” technology, which aims to integrate itself seamlessly into the home, assisting customers without intruding into their lives. The Google Home line of products, for instance, works by listening for the phrase “OK Google” or “Hey Google,” which prompts it to respond to verbal commands using natural language processing. Other companies, like Amazon and Apple, have developed products that work along the same lines, with the goal of becoming an essential part of people’s lives without making their presence obvious or intrusive. As the trend of integrating AI into the home continues, other manufacturers are likely to develop appliances that use technology to optimize the efficiency of life in the home.

Embed from Getty Images

Take, for example, LG, which is working on technology to improve the oftentimes-difficult experience of installing new appliances like washing machines and dryers, as well as technology that improves the customer service experience when consumers run into issues with the company’s products. So-called “smart appliances” work by integrating artificial intelligence deeply into all aspects of the appliances, helping users install appliances, detecting and communicating hardware errors, and even providing customer service using chatbots which are programmed to understand and respond to common consumer queries. LG’s latest iteration of washing machines and dryers connect wirelessly to customers’ smartphones using their proprietary ThinQ mobile app, which notifies users when the installation of appliances is completed and also provides users with information about the functioning of their devices as time goes on.

LG’s newest washing machine and dryer, called the LG TwinWash and ThinQ Dryer, include a number of sensors and artificial intelligence programs to streamline and improve the laundry experience. The TwinWash washing machine, for instance, includes voice recognition technology to allow users to operate the machine in a natural way without using buttons, and the washing machine can even give users verbal laundry advice depending on the types of stains on clothing. The machines also intelligently discern the softness of laundry in order to minimize fabric damage and improve washing quality. Additionally, when these appliances are released to the general public, users will be able to receive updates via their smartphones notifying them of problems with the devices that need to be addressed as well as reminders for scheduled maintenance in order to extend the life of the products and, in theory, reduce overall costs.

Embed from Getty Images

The appliances that work with LG’s ThinQ app are not limited to washers and dryers, but include everything from refrigerators, robotic vacuum cleaners, smart TVs, and more. LG’s InstaView smart refrigerator introduces a number of features that set it apart from today’s standard refrigerators, with the aim of saving energy and improving food freshness, among other features. The InstaView refrigerator is packed with a number of features that were once considered squarely in the realm of science fiction; for instance, the fridge includes a camera that films the inside of the refrigerator when the door is closed, which users can view on the device’s LCD touchscreen or remotely using the ThinQ app on their smartphones. Users of the InstaView refrigerator can also program the appliance to remind them when their food expires, and the device even includes Amazon Alexa, a popular voice assistant that can play music, check the weather, and even help users shop for groceries. The fridge also alerts users when the door is left open, produces large amounts of ice for parties or other occasions, and can enter a low-power mode that keeps food fresh when the user goes on vacation.

Clearly, such advanced home appliance technology is not for everyone, and consumers may reasonably question the usefulness of many of these products’ features. When LG’s line of smart appliances releases in the United States, they are likely to be very expensive, limiting their appeal to a small audience of consumers. However, if history is any indication, the technology that powers these appliances is likely to grow more sophisticated and cheaper with time, and it may just be a matter of time before smart appliances become a commonplace and even mundane sight in the home.

Colored Keyboard

Al Research the Games: Deepmind Beats Nearly AllComers

When DeepMind, Google’s AI research outfit, set out to demonstrate its latest breakthrough, it had to confront an added twist: how do you set your robot free to play games on the internet without anyone realising they’re competing against it?

The company caused a stir when it announced that its AlphaGo AI had beaten a world-class player at the ancient Asian board game Go. A few months later, it beat the world number one player.

But for the deeply strategic real-time war game StarCraft II, it had a different goal: to reach “grandmaster” standard – putting it in the top 200 players worldwide – on the game’s public servers, building its ranking the same way any human player would. That meant being matched with a steadily improving cadre of other human players, and winning against them consistently enough to be promoted.

StarCraft may seem like an odd next step, for a team that has previously taken on chess and Go, but the game has some qualities that make it interesting to researchers. It’s real-time, with millions of possible actions each second, and a vastly more complex roster than the six pieces of chess. Most importantly, it features hidden information: for the first few minutes of the game, it’s impossible to even see what your opponent is doing, let alone work out what they’re planning.

Embed from Getty Images

That means strategies have to be flexible enough to account for surprises, and need to incorporate mind-games as well. There’s also an advantage in a community where even the best players in the world can be found playing each other online, ranked according to a very public algorithm, with a ton of data flying around.

Players were told the new AI, dubbed AlphaStar, would be online, and were given the option of opting-in to play it. In order to ensure it achieved its rank fairly, it had to play its games anonymously, so that opponents didn’t spend more effort trying to trick it or break it than they did trying to win.

“There was a bit of a meme where people started asking ‘are you AlphaStar’ to others,” said DeepMind’s David Silver, one of the company’s co-founders and a lead author on the Nature paper announcing the StarCraft II victory. “We had the policy to just not chat – other than wishing people good luck, and then ‘good game’.”

The need to remain anonymous did also turn the experience from a test of raw skill into a sort of “Turing test for video games”, said Silver’s colleague Oriol Vinyals. “AlphaStar needed to play like a good human, not like a superhuman.”

Embed from Getty Images

That meant taking a different approach from previous StarCraft AIs, which tended to lean on the abilities that only a computer could have. In a game where human competitors track their “actions per minute”, a professional-level player may hit three or four hundred, while some AIs were acting thousands or tens of thousands of times over a sixty-second period. At other times, AIs were given near omniscience, with all the information available over the entire map plugged into their systems at once.

“We really wanted to have an interface that we believe was reasonable from a capability standpoint,” says Vinyals. “So we added this notion of a camera view, which is very crucial for players to control where in the map they’re actually focusing on, and we also reduced the peak actions per minute, to 22 actions in a span of five seconds.” In other words, the AI is forced to play much more like a human.

All of which is moot if the AI gives itself away by, well, playing like a robot. Luckily, it doesn’t – quite. In the first series of matches played publicly, in January, AlphaStar did exhibit one slightly mechanistic behaviour, falling prey to an almost cartoonish tactic where its opponent, the human player MaNa, moved a unit into and out of its field of view, changing its behaviour each time. It worked for MaNa to eke out the only win the humans scored over those first 11 matches.

More interestingly, the AI did develop its own understanding of the best tactical play, occasionally differing from the generally accepted practice among pros. The intricacies are a bit specialist, but reinforce the idea that simply teaching an AI to perform a task to human level can improve our understanding of the work itself.

“AlphaStar has been an amazing experience,” Oriol says. “Not because we beat most humans. But it’s more like that we were able to see what some limitations might be, to inspire research that will come, hopefully in the next few months or years and decades. Picking harder and harder problems and trying to be very good at them has been clearly the way so far.”

Doctor with Patient

Researchers Argue Sex and Gender Analysis Improves Science

It has long been understood in scientific circles that unconscious biases, particularly those relating to sex and gender, can have a negative impact on the objectivity of scientific findings. While the goal of science is to discover the truth in as objective a manner as possible, scientists are prone to the same unintentional, biased assumptions as anyone else, and the quality of scientific work can be affected. For instance, the appropriate dosage for a medicine may be devised with the assumption that the patient is male, leading to suboptimal dosage recommendations for women. As another example, safety equipment too can be designed with the physical concerns of men in mind, negatively affecting women who use the equipment. And as machine learning technologies advance, engineers are realizing that machine learning programs are capable of picking up on human beings’ unconscious biases and replicating them, perpetuating the problem.

In light of these realizations, much conversation has taken place regarding how best to correct for sex and gender bias in science. This concept is explored in an article posted in Nature entitled “Sex and gender analysis improves science and engineering.” The article’s authors argue that taking sex and gender into consideration while conducting science not only benefits less-advantaged individuals by recognizing the institutional challenges they face, but also improves the quality of science itself, as unconscious biases are identified and corrected. This approach, the authors claim, benefits multiple scientific fields, including medicine, artificial intelligence, and even climatology. 

Embed from Getty Images

While many consider the terms synonyms, the researchers explain the difference between sex and gender, defining the former as including mainly biological attributes, whereas they define the latter as “psychological, social and cultural factors that shape attitudes, behaviours, stereotypes, technologies, and knowledge.” This distinction is important because sex and gender interact in complex ways; for instance, there exist physiological differences relating to the experience of pain between the sexes, and gender impacts how patients communicate pain with doctors and researchers. The researchers point out several improvements which have been made in this area over the past several decades; for instance, crash test dummies were originally based on a male physique, but now represent more diverse body shapes, allowing engineers to design vehicles that are safe for a larger number of people. However, they also point out areas for future improvement. 

As advanced technology continues to influence society, ensuring that it doesn’t perpetuate harmful stereotypes takes on additional importance.

In their paper, the scientists focus on the surprising and complicated ways sex and gender manifest across a variety of disciplines, with the most focus placed on marine science, biomedicine, robotics, and artificial intelligence. The authors discuss how sex impacts science even in non-humans, as male and female marine life react differently to the effects of changing ocean temperatures, an observation which has generated insights about more accurately modelling the effects of climate change. In human beings, sex differences account for disparities in responses to various medicines, such as vasopressin and cancer immunotherapy, for biological reasons including differences in amounts of testosterone and estrogen and overall body composition.

Embed from Getty Images

Perhaps more surprisingly, artificial intelligence is a field in which unconscious biases can make their way into technologies, unintentionally perpetuating cultural biases and stereotypes. For instance, advertising algorithms are more likely to automatically serve ads for high-paying jobs to men than to women, and automatic image captioning algorithms tend to misidentify pictures of men in kitchens as women. As advanced technology continues to influence society, ensuring that artificial intelligence doesn’t perpetuate harmful stereotypes takes on additional importance.

The authors conclude by proposing solutions to many of the problems with sex and gender biases in science they identify. One suggestion is to foster greater interactions between the scientific community and the humanities, including social scientists. Allowing for interdepartmental conversations in this way helps scientists to learn about how biases emerge and affect human reasoning, and can incorporate this knowledge into their work. Additionally, the researchers advocate for greater transparency in scientists’ reporting by including variables relating to sex and gender in their data analyses. 


TikTok: The Future of Social Media?

The expansion of social media into our everyday lives seems to be unstoppable. Even for those of us who stubbornly refuse to sign up for platforms like Facebook, Twitter, and Instagram, these services influence us on a daily basis: Twitter has become the platform of choice for companies, celebrities, and politicians to give their immediate reactions to current events; and major events like parties and reunions are planned via Facebook, causing headaches and potential missed opportunities for those without an account. And as this expansion continues, new social networking services are popping up that re-write the rules of how social apps should operate, exploiting our collective human desire for connection and stimulation with increasingly sophisticated artificial intelligence programs. One such app is TikTok, which paves the way for the future of social media by doing away with the traditional framework of aggregating content from accounts selected by the user in exchange for pursuing engagement as its primary goal.

Unlike most popular social media services, TikTok was created not in the United States but in China, and not as an independent start-up but as the flagship product of an established artificial-intelligence and machine learning company called ByteDance. Whereas other platforms onboard new users by encouraging them to connect with the accounts of people they know, TikTok drops you straight into a never-ending stream of content, aggregated seemingly at random, before you even make an account. You’re not immediately invited to view the content created by people you know, as you might expect; rather, the app invites you to follow trends, organized by hashtags under the “Discover” tab, and create content befitting these trends to gather an audience, albeit an often-ephemeral one.

Embed from Getty Images

While the short videos you see in the app seem to be hand-picked for their humor, creativity, or charm, they are in fact curated by complex algorithms fed by data collected on the app’s massive user base. Over time, the app leverages machine learning and artificial intelligence to build a profile of your viewing and engagement habits, fueling an ever-more addictive and engaging stream of content tailored to your preferences. This shift away from content posted by your friends towards content the app predicts you’d enjoy is representative of a larger overall trend in the evolution of social media; as platforms like Facebook, Instagram, and Twitter have grown, their feeds began to organize and select content according to a set of rules defined by the company, rather than by the user.

This approach has proven to be an overwhelmingly successful one for ByteDance, so much so that concerns have been raised about the app’s addictive qualities, particularly among the community’s younger members. In response to these concerns, the company introduced a “Digital Wellbeing” section of the app’s Privacy and Settings menu, enabling users and their parents to set restrictions on how long the app can be used per day. Additionally, many have expressed worries about the app’s potential to show inappropriate or dangerous content to a young audience, and TikTok was fined for violating the Children’s Online Privacy Protection App in the US, leading the company to implement a kids-only mode which prevents children from uploading data to the app. These concerns, and the resulting updates to the app, represent the exceedingly few instances in which TikTok’s developers have taken a direct approach in managing the spread and proliferation of content, which is almost entirely directed by a combination of the community’s contributions and the algorithm’s attempts to proliferate attention-grabbing content.

Embed from Getty Images

TikTok represents a fundamental shift in the social app ecosystem, away from the goal of connecting users to one another and towards the goal of generating engagement for its own sake. All aspects of the app’s design, from its content-focused front-end to its leverage of sophisticated algorithms to curate content, are engineered to maximize engagement while remaining completely agnostic as to what type of content gains traction. This focus on engagement explains the app’s resounding success, especially among young people: in the modern social app ecosystem, developers compete with one another not for your money but for your attention, and as TikTok is singularly focused on maximizing its share of your attention, it sucks the life out of competing social apps. As such, one can expect other platforms to continue to gravitate towards this engagement-centered philosophy, and while the greater social ramifications of this approach are as of yet hard to predict, they are bound to be substantial.