Since there is no legal duty to delete this data, it accumulates over children’s lives and could last indefinitely.
Children frequently yell commands to Apple’s Siri or Amazon’s Alexa in busy houses all around the world.They might turn asking the voice-activated personal assistant (VAPA) the time or asking for a hit song into a game.There is much more going on than what may initially appear to be a routine aspect of family life.The term “eavesmining”, a combination of eavesdropping and datamining, refers to the continual listening, recording, and processing of acoustic events by the VAPAs.As audio traces of people’s life are datafied and examined by algorithms, this creates serious privacy and surveillance problems as well as discrimination concerns.
When we apply these worries to kids, they become more serious.Their lifetime data collection goes much beyond what was ever done for their parents, with far-reaching effects that we have only just begun to comprehend.
As it expands to include mobile phones, smart speakers, and an ever-growing variety of goods that are connected to the Internet, the adoption of VAPAs is happening at an astounding rate.Digital toys for kids, home security systems that listen for break-ins, and smart doorbells that can hear conversations on the street are some examples of this.
The gathering, storing, and analysis of sound data raises important issues that affect parents, young people, and children.Alarms have already been raised; in 2014, privacy activists expressed alarm about the volume of listening done by the Amazon Echo, the types of data being gathered, and the potential uses of the data by Amazon’s recommendation algorithms.
However, despite these worries, the use of VAPAs and other eavesdropping technologies has grown rapidly.
According to recent market analysis, there will be more than 8.4 billion voice-activated devices worldwide by 2024.
Recording more than just speech
As VAPAs and other eavesdropping devices overhear personal aspects of voices that unintentionally expose biological and behavioural variables like age, gender, health, intoxication, and personality, more information is being captured than just spoken words.
“Auditory scene analysis” can also be used to gather information about acoustic surroundings (such as a noisy apartment) or specific sonic occurrences (such as broken glass) in order to develop assumptions about what is happening there.
Eavesdropping devices have a history of working with law enforcement and being subpoenaed for information during criminal investigations.This prompts worries about the creeping surveillance and family and child profiling in different ways.
Smart speaker data might be used, for instance, to generate profiles for “troubled youth”, “disciplined parenting approaches” or “noisy families”.Future governments might utilise this to create risky profiles of people who depend on social aid or families in distress.
Also being promoted as a way to keep kids safe are new listening devices called “aggression detectors”.These dubious technologies, which consist of microphone systems loaded with machine learning software, assert that they may help predict violent situations by listening for sounds like glass breaking as well as vocal cues like volume increase and mood.
At law enforcement conferences and in school safety periodicals, aggression detectors are advertised.Under the pretence of being able to prevent and identify mass shootings and other instances of lethal violence, they have been placed in public areas, hospitals, and high schools.
But there are significant problems with these systems’ effectiveness and dependability.One particular model of detector frequently mistook children’s verbal cues, such as coughing, yelling, and cheering, for signs of aggressiveness.This raises the issue of whose safety is being protected and whose safety will be compromised by its design.
The interests of all families won’t be uniformly protected or served by this type of securitized listening, and certain children and teens may suffer disproportionate harm.Voice-activated technology is frequently criticised for reproducing racial and cultural prejudices by enforcing vocal conventions and incorrectly classifying culturally diverse forms of speech in terms of language, accent, dialect, and slang.
We should expect that racialized children and young people’s words and voices will be disproportionately misunderstood as seeming aggressive.Given the deeply ingrained colonial and white supremacist traditions that persistently enforce a “sonic colour line”, this unsettling prognosis should not be shocking.A good policySince children’s and families’ sonic activities have become important sources of data that may be gathered, monitored, saved, processed, and sold to thousands of third parties without the subject’s awareness, eavesdropping is a rich source of information and surveillance.These businesses are profit-driven and have limited moral responsibilities to protect children’s privacy and data.
Since there is no legal duty to delete this data, it accumulates over children’s lives and could last indefinitely.It is unclear how long and how far these digital footprints will follow kids as they grow up, how widely this information will be disseminated, or how much it will be compared to other information.These issues have significant effects on children’s life now and in the future.
Eavesdropping poses a wide range of hazards to privacy, monitoring, and prejudice.Individualized solutions, including informational privacy education and digital literacy training, will be ineffectual in resolving these issues and put too much of the on on families to acquire the literacy required to prevent eavesdropping in both public and private areas.
The development of a broad framework to address the particular dangers and realities of eavesdropping must be taken into consideration.The creation of Fair Listening Practice Principles, an auditory version of the “Fair Information Practice Principles”, might aid in assessing the platforms and procedures that have an effect on children’s and families’ access to sound.