Understand the basics of how computers comprehend and respond to human language. Alexa, play my favorite song.’ You can utter these five words without moving an inch and within seconds, your favorite song will start playing. But have you ever wondered how Alexa, your virtual assistant, [...]
How voice tech has advanced, and its impacts on business and humanity.
When Siri was launched in 2012, the world was thrilled and disappointed. The fact that a virtual creation – little more than a soothing voice on a smartphone speaker – was capable of taking and acting on verbal commands, felt like a sci-fi movie come true. Yet, Siri’s performance back then was far from satisfactory, as it struggled to understand human instructions properly.
However, technology has advanced since. The world has witnessed a huge leap in Artificial Intelligence (AI) technology in the past decade. Voice tech, benefiting from progress in AI development, has improved enormously and it is predicted to be the dominant user interface in the future.
Proliferating voice tech: how does it thrive?
The major barrier to voice tech development is its reliance on the ability to decipher human speech. Speech recognition can be complex due to the dynamic nature of languages. It requires us to combine multiple elements – tones, pitch, slang, wordplay – to understand the true meaning of a sentence. Communicating like this is an ability learned over millennia of evolution – small wonder that machines might not be able to pick it up at once.
However, AI has changed the game. Thanks to advances in psychology and computer science, engineers can now decode the mechanism that dictates human decision-making and thereby generate corresponding formulas that predict human behaviors. AlphaGo, a computer program that plays board game Go and outperforms its human counterpart, is a vivid example of AI advancement.
Voice tech, benefiting hugely from the advancement in AI, has to resolve the difficulties that used to stymie its proliferation. As early as late 2017, Google announced that it had achieved 95% word accuracy in speech recognition for U.S. English.
The technology has also evolved to go beyond mere speech-to-text recognition: Alexa, another voice assistant developed by Amazon, is capable of adapting to users’ language overtime to accommodate unique accents.
As more voice-enabled products and services roll out to the market, the voice tech industry is anticipated to become a very lucrative business indeed. According to a 2015 Grand View Research report, the voice tech market is estimated to be worth nearly US$32 billion by 2025. Voice tech is now being widely used in assisting customer services, restaurants are offering voice ordering, and banks have introduced voice as a form of biometric identification to combat fraud.
Apart from benefiting businesses, voice tech has also brought positive change to the healthcare sector by increasing accessibility. Tel Aviv-based Voiceitt, for instance, develops algorithms that assist people with speech impairments.
The Voiceitt program acts as an interpreter to other voice assistants that are normally unable to figure out what the user is saying. It can also be used in therapy and medical appointments to allow doctors to understand patients with linguistic impairments.
Trading privacy for convenience
As beneficial as voice tech might seem, there are still plenty of lingering concerns over its applications. Among them, the biggest one is privacy.
The very essence of voice tech is to allow people to simplify processes by using verbal commands instead of physically typing on a device. People can save themselves the trouble of searching for nearby restaurants by shouting across the room to a voice assistant, who will then read out all the recommendations for them.
However, this also implies that a voice-enabled device needs to standby and operate continuously to make instant responses. This implies that users are constantly monitored, leading to a more vigorous collection of personal data.
Some voice-enabled devices allow users to turn on the voice detection function manually: for example, Apple provides an opt-out mechanism for automatic voice annotation and allows users to delete voice dictation history. However, Apple didn’t always provide an opt-out, and changing these settings wasn’t exactly straightforward.
It ultimately took years and a series of audio leaks that stirred up substantial trouble for Big Tech before the major companies investing in voice assistant technology changed things up. Most recently, on August 5, Google automatically switched all users who have ever utilized Google’s AI to opt out of storing audio recordings. Users who wish to use the feature will be required to opt back in.
With constant monitoring intertwining with steadily-increasing usage of voice tech devices, sooner or later, your voice assistant might know you better than you do. Big Tech companies will know – if they don’t already – your daily routines, your preferred timing for ecommerce deliveries, your opinion on environmental issues, and more. All these allow their algorithms to build astonishingly accurate profiles of their users and make tailored recommendations to them.
On an even more futuristic note, voice tech might become capable of reading people’s emotions. It might sound like a pipe dream right now, but tech giants like Amazon and Google have already filed patents for technology to read emotions in people’s voices, which would give algorithms a glance into mood stability and mentality in a bid to provide services tailored to an even greater degree. For instance, if your voice assistant detects distress or weariness in your voice, it might recommend that you order pizza instead of making dinner.
Recent scandals of tech titans misusing personal data have already sparked widespread concerns over data security. In a 2017 Digital Rights in Australia report, 62% of 1,600 participants felt they weren’t in control of their online privacy, and about 47% were concerned that the government could violate their privacy. New regulations on data protection generally lag behind the explosive pace of development in technology.
The enforcement of data protection law is even more difficult, as there is a fine line between legitimate and illegitimate collection of data. Multiple considerations need to be taken into account, such as the context of data, the purpose of collection, the methodology used in the acquisition of users’ consent, and region-specific laws. All these requirements make it costly and timely to enforce data protection laws.
Moreover, the cost of breaching customers’ privacy is significantly lower than the profits it can generate. Facebook was fined US$5 billion for mishandling users’ personal information in 2019, the largest fine ever imposed on any company for violation of customers’ data. Nevertheless, the amount is still small compared to Facebook’s total annual revenue, which was around US$22 billion in 2018.
Fines for data breaches vary across regions, but Big Tech has been making such unrestrained profits that any fine is likely to be an insufficient deterrent. It’s not hard to imagine that tech companies are willing to take the risk for a chance at higher margins.
The silver lining
The rise of voice tech has helped Big Tech to sneak into new avenues of personal data collection – seemingly an inevitable trend in this technological era. Users of virtual home assistants like Alexa might regard this as an unavoidable compromise for a more convenient life, but with growing awareness around personal data collection, there is a growing population of those who would prefer to steer clear entirely.
That said, there is still some hope for the future. As privacy and personal data safety gain more attention globally, legislative reforms have been carried out in various countries, re-examining outdated regulations and replacing them with more robust measures to ensure personal data security.
Most significantly, the EU has taken the leading role in controlling data collection and usage by implementing the General Data Protection Regulation in 2018, to address cross-border transfers of personal data and impose more responsibility and accountability on personal data collectors. Although the effectiveness of the regulation is to be examined as time passes, it is already creating a ripple effect that is encouraging reforms in this long-neglected domain.
Tech companies and businesses can also take active measures to reassure their clients and boost confidence in their data privacy measures, such as providing ‘opt-in’ systems like that of Google. Companies could also be more open and honest about their use of collected data, allowing consumers to choose whether the added convenience is worth giving up personal information..
It can be bewildering to live in such a rapidly digitizing era, with technology granting us enormous benefits while also posing unprecedented threats. But with a growing dialog on privacy protection and the initiation of gradual reforms, eventually, we will find a way to achieve the delicate balance between convenience and privacy.