Date | May 2021 | Marks available | 6 | Reference code | 21M.1.SL.TZ0.3 |
Level | SL | Paper | 1 | Time zone | no time zone |
Command term | Analyse | Question number | 3 | Adapted from | N/A |
Question
Voice simulation
Lyrebird voice-cloning software was created by three university students from Montreal. This technology was originally developed to change the lives of people who have lost their voice. These people can use the software to create a new voice by choosing one from samples provided in the software.
However, Lyrebird can also clone voices. Lyrebird claims that it can clone a person’s voice in just 60 seconds. Users only need to access the website, create an account, and generate a voice print by repeating a number of predetermined sentences. Although the cloning is not perfect, the developers of Lyrebird believe that both the time to clone the voice and the quality of the cloning will improve in the future.
The software allows the user to alter the speech and add emotion (such as, anger, elation, and surprise), increasing the ways in which it can be used.
The initial voice print for the cloning process requires the most time and a lot of processing power. However, creating speech from the voice print is a simple process, generating around 1000 sentences almost instantly. The voice print is saved as a compressed audio file.
Lyrebird has considered the ethical issues regarding biometrics* and acknowledges that it highlights some key societal issues. Lyrebird has taken the decision to make the software available to all in an attempt to address these issues.
* biometrics: a means of personal identification through analysis of unique physical traits or behavioural characteristics (for example, through voice patterns or fingerprints)
Lyrebird uses a person’s voice print to determine their identity.
Identify three other methods of biometric identification that could be used to determine a person’s identity.
A person’s voice print can be compressed.
Identify three problems that may result from compressing the voice print.
Analyse whether it would be appropriate for Lyrebird to release the biometric data of a user to a third party.
To what extent do the benefits of using Lyrebird’s voice-cloning software outweigh the concerns that may arise from its use?
Markscheme
Answers may include:
- Fingerprints
- Eye/retinal scans
- Palm prints/hand geometry
- Facial recognition
Award [1] for each method of biometric feature identified up to [3] marks.
Answers may include:
- Possible quality loss from the original file when decompressed.
- Potential for file corruption in the compression process.
- The compression may use significant processing resources while it is taking place, leading to other programs running more slowly.
Award [1] for each problem that may result from compressing the voice print identified up to [3] marks.
Answers may include:
It would be appropriate:
- to law enforcement agencies for a warrant, subpoena, or court order
- in response to requests by regulatory, legal or government agencies, as the information cannot be obtained by any other means
- by withholding this information it would be working against the ethical principles of Lyrebird
- if it is necessary or appropriate to do so to protect our rights, property, or safety
- where the user has provided Lyrebird with written consent to disclosure of your biometric data.
It would not be appropriate:
- where the releasing of the data would lead to harm being caused to the person whose data is being released, for example, data that is linked to an underlying medical condition (values, ethics)
- where it is in direct contradiction of the policies that Lyrebird have for the sharing of data
- as it could be harmful to the reputation of Lyrebird
- as it is an invasion of privacy to release personal data without consent
- it is illegal under the laws of the country in which Lyrebird operates (e.g., GDPR).
Keywords: legality, regulation, policies, privacy, anonymity, security, transparency, ownership, access, sharing, consent, data, GDPR, power, ethics, values
Answers may include:
Benefits may include:
- It could help people who have lost their voice.
- It could create better voice-activated interfaces (systems).
- Characters in video games could have individual voices or could clone a player’s voice as a playable character (expression, identity).
- It could provide interactive training and learning.
- It could be used in educational software to produce a more realistic experience (systems).
- Any device that requires speech could be enhanced by more natural voices (appointment reminders, personal assistant applications, audiobooks, etc.) (systems).
- Parents could use the software to “read” to their children (relationships).
Concerns may include:
- It could be used to create fake news or weaken the public’s confidence/trust in news of all kinds.
- An increase in robocalls with realistic voices – scammers will be able to deceive people more easily or make fake calls to the police (values).
- It may lead to systems that use voice as a method of authentication becoming compromised, as scammers are able to quickly create a user’s voice print and access the system.
In part (c) of this question it is expected there will be a balance between the terminology related to digital systems and the terminology related to social and ethical impacts.
Keywords: health, privacy, anonymity, security, transparency, ownership, access, authentication, sharing, expression, identity, power, values
Refer to SL/HL paper 1, part c markbands when awarding marks. These can be found under the "Your tests" tab > supplemental materials > Digital society markbands and guidance document.