There is no doubt that AI is the hot word of the last period.
She writes poems and articles, draws and solves tests, and with each passing day we discover another new ability or trait.
One of the uses that is gaining momentum recently is the Deepfake and the Voice Generator.
The FAKEYOU website, for example, allows you to choose the celebrity you want to "imitate".
After you type the text, click on SPEAK, and the website will generate an audio clip for you in the voice of the person you selected.
The site will also allow you to animate the text on an image that you choose from the variety of images and gifs on the site: Donald Trump giving a speech, Walter White angry, Albert Einstein with a thoughtful look, etc.
This is just one example among many of the tools that exist today.
This video sums it up well:
When technology gets out of control
When Alfred Nobel invented dynamite, he did not imagine that it would be used to kill and fight.
Similarly, the exciting new features of AI also have dangerous applications that deviate from the original plan, which is, by and large, to make our lives easier.
One of the most disgusting uses is faking erotic videos of celebrities, a phenomenon that began with the arrival of technology to maturity.
While it makes sense to transfer the field of porn into the hands of machines, thereby preventing phenomena of exploitation and physical harm to young women, 'undressing' real women in photos and videos using AI and publishing the products without their consent is a serious phenomenon, and many porn sites, including Pornhub, have already announced that they will block such videos.
Thus, the technology that could have turned the bowl upside down forced the porn industry to do without it and stick to its old ways, at least for now.
good to know (in advance)
The secret to improving your sex life - now in a special sale
Served on behalf of "Gabra"
Another appalling use is in the field of fraud and extortion.
An investigation published in the Washington Post in early March revealed how scammers used artificial intelligence to mimic the voices of victims' family members.
The same crooks pretended to be a family member in need and managed to extract large sums of money from the victims, who rushed to help their relatives.
But if you thought it couldn't be lower than that, think for a moment about the phrase "mind engineering", and let's continue.
A step up in the quantity and "quality" of fake news
In the wrong hands, deepfake software becomes fake news poison machines.
Just a year ago, Russia released a fake video of President Zhelinsky calling on Ukrainians to submit to the Russian invasion, while on the other side of the border, a fake video of Putin announcing peace between Russia and Ukraine was also released.
If until now we thought it was difficult to follow fake news because anyone can write whatever they want online, now everyone can use whatever voice or image they want to say it!
Is there a way to deal with this?
In two words, it's complex.
Similar to features like OpenAI's AI Text Classifier, which can diagnose whether a certain text is "authentic" or if it was written by artificial intelligence, new tools like the Deepware Scanner are popping up like mushrooms after the rain with the goal of allowing users to diagnose deepfake videos, which are generally perceived as credible More because "there is nothing like a mirror".
Alongside this, a variety of methods are already appearing on the net to bypass the AI detector by changing the original language in which the text is written, or requesting in advance from the software to write in a more 'human' manner.
This means that we may also see an 'arms race' in the field of deepfake: more and more truth detectors will be developed, and in response more and more tools and methods will be developed to bypass those detectors and lend credibility to fakes.
The governments of the world are reacting with drowsiness
In October 2022, the State of Israel published a policy document that should be the cornerstone of appropriate regulation that will follow.
The document emphasizes the importance of principles such as maintaining privacy, respecting the rule of law and the public interest, and more.
Later, the Italian government announced a ban on the use of ChatGPT, OpenA's text generator, in a move that most illustrated the government's lack of understanding of the situation, since it is only one of many tools that are expected to disrupt entire industries from the ground up.
At the same time, experts are calling for a halt to the development of AI models for a period of six months.
Hundreds of businessmen and academics have already signed the petition, including Elon Musk, who owns Twitter, which recently got rid of its AI ethics team.
Even in the United Kingdom, they did not remain indifferent and on March 29, the British government published a report on artificial intelligence with the aim of raising public confidence in artificial intelligence and encouraging its use for innovation at work.
The report presents five binding principles for those who use artificial intelligence: safety, transparency, fairness, responsibility and correction of harmful results, if any.
According to the report, companies that use AI should provide an explanation of how it is used and the decision-making process, while referring to the risks arising from its use.
tame the monster
It is not yet clear what the impact of those reports, policy documents and petitions calling for a halt to the gallop to an artificial world will be, but what is certain is that AI-based deepfake tools are already fundamentally changing the rules of the game.
The real question is whether humanity, which is currently busy signing beautiful statements, will actually succeed in overcoming the 'monster' it created, or are we at the beginning of an era where it is really impossible to distinguish between lies and truth?
Attorney Reot Goldman is the CEO of the Geometrics Group, a member of the "Helsinki" committee at Belinson Hospital, and a member of the board of trustees of Schneider Children's Hospital
Marketing and digital
in the headlines
Tags
Fake news
artificial intelligence
transparency
/cloudfront-eu-central-1.images.arcpublishing.com/prisa/CY2KB3TF3FGHVCUFL5I5IE3W3I.jpg)
