What are the ethical and legal considerations for using AI-generated audio and text-based content? originally appeared quora: A place to gain and share knowledge, helping people learn from others and better understand the world.
answer By Matt Hocking, co-founder of WellSaid Labs. quora:
In approaching responsible AI, we built the foundation of WellSaid Labs on four key pillars that are generally applicable to all AI companies.
Accountability – As a company developing machine learning models, you need to fully understand the process of AI learning, or “AI understandability,” or why an AI system reaches a particular decision or output. Owning and understanding the entire machine learning process, as opposed to licensing it from another company, reduces undesirable outcomes such as biased or factually incorrect answers, and helps you reduce the risk of undesirable outcomes such as biased or factually incorrect answers when they occur. can be dealt with. After all, artificial intelligence is a technology, and it can be made worse or better depending on the intentions of the humans using it. Feigning ignorance or refusing to engage with the fundamental processes behind AI in order to avoid liability, save costs, or cut corners is irresponsible and dangerous.
Transparency – We always make sure that voice actors understand the implications of lending their likeness and vocal identity to an AI voice before engaging with us. Although this process may seem unique to her AI voice, anyone lending data to a machine learning model must give informed consent before lending it. We provide support for specific use cases within the terms of our Terms of Use, while making users aware of the associated risks associated with participation.
Privacy and security – This is one of the most widely applicable responsible AI practices that should be mandatory for all companies using machine learning. AI models are often trained on data from real people, and that data must be protected regardless of its use. At Wellsaid Labs, we have built and maintain systems and practices that ensure the security and privacy of our users' and voice actors' data.
fairness – Remove voice actors from our platform if necessary (including the ability to create voices from voice avatars of voice actors across all users). This is something we have done in the past. People who have their likeness used by AI should have the option to opt out at any time. AI has progressed to the point where it is difficult to parse actual images and sounds from images and sounds created by AI, so to create AI sounds and images, he is the person who uses the ID. It has become important for people to continue to agree. Our voices are not retained forever, and voice talent choose to work with us voluntarily. Although they are not bound by any contractual obligation to work with us for a specific period of time, many choose to partner with us for years.
this question originally appeared quora – A place to gain and share knowledge, enabling people to learn from others and better understand the world.
follow me twitter. check out my website.