What challenges does Amazon Alexa upgrade to AI voice synthesis?
arts/
According to the news on June 23, Amazon is ready to let the Alexa voice assistant simulate anyone's voice, and then control the home's smart home system through the Echo speaker.
Behind this technology is based on artificial intelligence, that is to say, as long as we speak to Amazon's Alexa voice assistant, the system can establish recognition based on our uploaded voice, and can also synthesize and simulate our voice based on our voice. vocals.
To demonstrate the maturity and applicability of this technology, Amazon also demonstrated new features at the conference. In a video, a child asks Alexa if she can read "The Wizard of Oz" in her grandmother's voice. Alexa confirms the command, modifies the voice, and then reads it at a gentle pace that sounds like a real person and doesn't feel mechanical.
Rohit Prasad, Amazon's senior vice president, said the new features could keep memories going during the pandemic when many people have lost loved ones. Another Amazon executive revealed that Alexa currently has 100 million customers worldwide, a number that is in line with Alexa device sales. In January 2019, Alexa devices officially went on sale.
The release of this technology by Amazon also brings us three very important considerations:
First, the judicial challenges brought by artificial intelligence technology, that is, when artificial intelligence can imitate human voices, it will also bring challenges to the current judicial system . It is mainly reflected in how to deal with malicious evidence of synthetic speech. For example, some people use artificial intelligence's real voice synthesis system for improper purposes, and then claim some fake real voice, and submit it to the judicial system as evidence, which is very important to us. The challenges brought about by the judicial system will inevitably increase the workload and difficulty of judicial technical identification of voice evidence;
Moreover, when the speech synthesis of artificial intelligence reaches a certain level, that is, when ordinary technology cannot recognize it, it will bring great trouble to the justice of justice. Especially in the case of re-recording based on artificial intelligence "real people" imitating speech, it will bring great challenges to the current judicial system.
Secondly, the fraud risk brought by artificial intelligence technology is when these artificial intelligence-synthesized "real people" voices are used in telecommunications fraud, such as the situation in which Xuetong data was sold in the overseas data black market in the past two days.
When the personal privacy data of some college students is sold, if there is still some voice information in the data, or the conversation voice information of related personnel. These voice dialogue information is obtained by fraud gangs. They use big data information to sort out the information of the fraud target, then find the points that can be defrauded, and then use artificial intelligence to synthesize voice to talk to the fraud target.
At this time, it is very easy for the general public to fall into the trap of fraud, especially for those who do not pay attention to cutting-edge technology fields . Of course, from this point of view, you must pay attention to Geinge's column, at least it can help you understand some traps and crises under the new technology.
Finally, the security risks brought by artificial intelligence are mainly in two aspects. On the one hand, it is aimed at some current systems and operations based on voice recognition control, especially systems and devices that use personal voice as control and recognition operations, obtained on the Internet cloud, or premeditatedly obtained some specific objects. After the real voice. Then use the artificial intelligence "real person" voice synthesis system to synthesize, and then use the synthesized voice to control these systems and equipment, which will cause great security risks;
On the other hand, it is not a good choice for smart homes. After we intercept the user's voice, we use this system to synthesize, and then use hacking technology to open the user's control port and convert these control voices. upload it. The haunted events in the movie may become possible. By controlling the lights of the home and the sound to play the horror lights and sounds, it will bring great trouble to normal human life.
Therefore, in the face of the current increasingly mature and intelligent artificial intelligence speech synthesis technology, in the face of the deepening application of smart devices, security has become a major issue for technology companies and legal supervision.