2023 seems to be the year when AI will truly come into the public consciousness in a significant way. We have already seen how ChatGPT is generating conversation – both positive and negative – while also pointing to a future of AI integration. Microsoft is also a major investor in AI and the company's new VALL-E could transform AI speech accuracy and quality.
VALL-E is essentially an AI synthesis tool that aims to remove the uncanny valley from AI speech while also being able to mimic human speech. Microsoft used 60,000 hours of English speech information to train the AI and is now showing the results in a research paper in collaboration with Cornell University.
AI is already able to synthesize human speech and while it is very advanced, being able to avoid uncanny valley has proved a major obstacle. This is where a human can detect something is not right with the AI. For example, the human speech from the AI sounds real but people know it is coming from a machine.
In some ways, it is arguable that the uncanny valley should always remain as some sort of failsafe in AI. Microsoft and other developers think not, and VALL-E is an example of how the company is breaking through the uncanny barrier.
A demonstration of VALL-E on GitHub shows audio sampler that go through a range of quality to unnatural to near perfection. The system also need very little input to produce convincing results and may one day be able to learn how to mimic human voices.
Controversy
While VALL-E is not publicly available, it or a predecessor one day will be. When that happens, the conversation on the dangers of AI will continue. Either the technology is going to transform the way we live or it is going to destroy civilization and plunge us into a Terminator-style war with machines.
No matter what side of that fence you are on, tech companies are continuing development either way. A demo of the ChatGPT chatbot was released in late November and is available from OpenAI here (log in required). ChatGPT can provide accurate answers to trivia questions, while also being able to generate AI content such as poems and songs.
Another feature of the AI is the ability to help debug code, which could make it an ideal companion for GitHub Copilot. However, there have been plenty of reports that ChatGPT often gives wrong answers that look correct. This confusion led to Stack Overflow temporarily banning the chatbot last month.
ChatGPT is just the start, and it will improve through each iteration. The era of AI permeating our lives is here to stay. Microsoft's VALL-E and even technology such as deepfakes suggest a future where AI can literally mimic a person is not too far away.
Tip of the day: File History is a Windows back up feature that saves each version of files in the Documents, Pictures, Videos, Desktop, and Offline OneDrive folders. Though its name implies a primary focus on version control, you can actually use it as a fully-fledged backup tool for your important documents.
