If we were told that we were being bugged and wiretapped 24/7, 365, all of us would go into fever fits. The fear that our lives are being monitored every hour, every minute and every second is not a truth that we are willing to hear or believe in, for that matter. But here is a piece of information that will hit us like a freight train – every user of voice-activated tech (which is almost all of us) is always under surveillance. The microphone that the technology uses is constantly listening in on us and recording because it is on alert about when and where it will hear the word “on” so that it can do as commanded. This implies that the device might continue to capture conversations in your home or workplace even when it is not actively in use. Maybe in our subconscious, we were already aware of this, but decided to brush it off because after all what harm can it do? However, what if this is only one-hundredth of the detriments that voice-activated tech was responsible for? What if there is more? Will we still turn a blind eye?
Voice-activated tech is intriguing and equally concerning
The speech and voice recognition market is expected to achieve a value of $56.07 billion by 2030, with a compound annual growth rate (CAGR) of 19.1% from 2023 to 2030.
The market’s expansion is fueled by the increasing adoption of voice biometrics, the incorporation of voice-enabled devices into automobile infotainment systems, and the widespread availability of voice-activated devices. While this will open so many doors for the world of technology and humanity in general, there is not one but many burning issues that make its developments concerning.
One of the biggest worries that ensues voice-activated tech is the fact that it may not be secure. The right to privacy is snatched away from its user. This is because once it has access to the user’s voice and personal data, misusing it for other purposes including targeted advertising is a piece of cake. It is possible that the minute that the voice is recorded and stored, data could be shared with third-party companies. The collection of voice data also supports personalised marketing efforts by utilising insights gathered from these recordings. There’s a concern about potential data breaches that could expose individuals to cybercriminals attempting to mimic their voices.
In an article on ABC News, the writer, Max Zahn quotes Marc Rotenberg, the founder and executive director of the nonprofit Center for AI and Digital Policy. He emphasised the growing issue surrounding the accumulation of voice recordings, especially as more people embrace voice-activated devices like Alexa and Siri. He referred to it as a looming challenge with significant implications.
At the beginning of 2023, Whole Foods, a grocery chain owned by Amazon, had reached a settlement to pay nearly $300,000 to its employees. This settlement addresses allegations that a voice-assisted tool used to monitor worker productivity in a Chicago warehouse had recorded the voices of employees without obtaining their consent.
Critics express concerns that voice-assisted devices may gather more sensitive data than users are aware of. This data could be exploited by companies for targeted advertising or the sale of personal information.
According to experts, an individual’s voice can potentially unveil a variety of personal details, such as height, weight, ethnicity, personality traits, and even potential health-related information. This revelation is based on discussions with scientists regarding the collection of voice data, as noted by Joseph Turow; a professor at the University of Pennsylvania’s Annenberg School for Communication, who explored the topic in his book; “The Voice Catchers: How Marketers Listen In To Exploit Your Emotions, Your Privacy and Your Wallet”.
One technology always leads to another. It is like a chain reaction. The existence of voice-activated technology will automatically result in voice cloning. That is just the way of the world. Voice cloning, a technology leveraging machine learning, is now accessible through various apps and websites, some of which are either free or have a minimal cost. Creating a voice model only requires a few recordings of an individual. Although the quality of the generated voice improves with high-quality recordings, individuals who have public recordings of themselves on social media or have been recorded in other settings may be at risk of having their voices replicated. Imagine this happening in a voice-based banking setting.
Voice-based banking is an industry that is growing at a rapid speed. As mentioned in Finextra, according to the Allied Market Research that was conducted in November 2022, the industry is expected to reach $3.7 billion by the year 2031. However, the entrance of voice-cloning technology into the picture brought forth a fair amount of scepticism. For instance, Microsoft’s VALL-E claims that voice cloning technology can replicate any voice using just 2-3 seconds of an original recording. VALL-E can also generate additional words from that snippet. It has the capacity to work with text in 23 different languages and grasp the context and essence of a sentence, going beyond mere word-for-word translation. While this may seem unbelievable, if it holds true, it could have a profound impact on voice-based remote communications.
Voice cloning has the potential to generate fake audio clips or voice commands that closely mimic an individual’s original voice, leading to concerns about potential misuse. This can give birth to long-term threats in the FinTech sector.
The security of voice-enabled remote banking systems is becoming a significant concern for both financial institutions that offer this feature and the customers whose financial information could be compromised. The idea of someone cloning your voice and interacting with your bank, such as checking your account balance or transferring funds to an external account, poses serious risks to voice-based interfaces worldwide.
Technology is travelling at the speed of light. Similar to how women are expected to be cautious in the face of misogyny and patriarchy and be ready with a manual on how to protect themselves, avid users of technology are met with the same fate. Individuals using voice-activated technologies should remain vigilant regarding critical privacy and security concerns. These devices accumulate substantial amounts of data, including voice recordings and personal data, which could potentially become targets for hackers and unauthorised access. To enhance security, it’s advisable to employ robust, unique passwords and activate two-factor authentication measures.
While reining in technology’s innovations is not a choice, taking precautions could be. In essence, it is vital to familiarise oneself with the privacy policies and data handling practises of voice-activated devices and, if the need arises, to implement measures to safeguard personal information.
(Sandunlekha Ekanayake)