Roger Quiles on LinkedIn: Fake Corinna Kopf accounts are using AI to convince fans she’s marrying… (2024)

Counsel Yaél Weitz was quoted in NPR in an article discussing the development of new tools to facilitate the detection of AI bot deepfakes and to make it more difficult for AI systems to create them. The article also discusses potential legislation to address liability for the use of people's likenesses without consent. Read more here.#ArtLawhttps://lnkd.in/g3c3vHzm

8

The current state of deepfakes vs. the government. Some misuse #ai for all the wrong reasons and how the government has been arming itself to protect people are summarized in this article. Again, I believe that #ai is a wonderful tool if used for the right purposes and in the right manners.#advertisem*nt #intellectualproperty #internetsecurity

Fake AI images are getting more sophisticated, but so are the ways to spot them. Learn how to tell the difference and avoid falling for digital trickery.

22

1 Comment

As AI gets more sophisticated, along comes the issue of trying to work out what is real and what is fake. In the below article you can learn more about how to spot fake digital images#ai #spotthedifference

1

The speed of AI adoption, concern and conversation is unlike anything I've ever been monitoring for a brand in culture. A nudge to look beyond the practical advancements you're being bombarded with to seek out those questioning the ethical implications and guardrails we need to address. A great space to watch is the conversation around the value of our voice [from the upcoming AI voice generated Edith Piaf film to the "No Fakes Act"bipartisan bill that "would apply throughout a person’s lifetime and, for their estate, 70 years after their death"]. Celebrities [and their families, a la Robin Williams's daughter] and one research team think they may have a protection solution - meet [WIP] AntiFake designed to protect against Deepfake abuse. "AntiFake scrambles the audio signal so that it confuses the AI model. The modified track still sounds normal to the human ear, but it sounds messed up to the system, making it hard for it to create a clean-sounding voice clone" There is no one "right road" and for every yay, someone will say nay, but simply put look too all the players, especially the ones trying to do the protecting. https://lnkd.in/dv9-RkXa

11

Congress Is Trying to Stop AI Nudes and Deepfake Scams Because Celebrities Are Mad: Lawmakers are proposing bills to safeguard celebrities and the public from AI-generated fraud. A surge in suspicious videos on platforms like TikTok, featuring famous personalities endorsing dubious products, has prompted this action. The proposed measures aim to combat the misuse of AI technology in misleading promotions affecting both celebrities and the general population. - Artificial Intelligence topics! #ai #artificialintelligence #intelligenzaartificiale

Deepfake AI images create new nightmares. Here are 5 ways we can cope | Fox News: Artificial intelligence deepfakes, which involve the creation and manipulation of images, are becoming a concern in society. Victims, especially women, are at risk of having false and sexualized images shared online. Legal and copyright issues arise as technology allows for the reproduction of deceased celebrities. Deepfakes also pose threats to political accountability and the spread of false information during elections. Government regulations and individual responsibility in sharing and vetting information are suggested solutions. - Artificial Intelligence topics! #ai #artificialintelligence #intelligenzaartificiale

There are 284 uncensored open-source models on Huggingface as of today. I've encountered them often but never really paid close attention to what they are and what purpose they serve. In this post, I will delve into these questions.❓ What are uncensored models?Uncensored models refer to language models that do not have built-in refusals or guardrails to prevent them from generating inappropriate, unethical, or illegal content. 🔞 Why uncensor a model?The primary argument for uncensoring a model is to cater to a wider audience with diverse beliefs and viewpoints. The alignment or censorship in models is often influenced by the cultural, legal, and political biases of the region where the model was developed. ☠ Dangers of uncensored modelsAs of now, anyone can download an uncensored model and use it as they deem fit. This unrestricted access opens up a Pandora's box of potential misuse, from generating inappropriate content to perpetuating harmful biases.IMO, while uncensored models have many uses, there needs to be some checks and balances to this. But how?Follow me for more updates on AI!#GenAI #AI #OpenSource #HuggingFace #LLMs #UncensoredModels

11

2 Comments

The dystopia 😱 of AI, I hope, ❌ never comes to pass.Currently, perfectly predicting human behaviour is essentially a black box, even if you could mind read. 🎉 Which secures the field of user research 🔍 from a total AI takeover. But if human behaviour could be perfectly predicted by some miracle, it would spell the 🚨 end of privacy.Marketers, Advertisers and, worse, scammers would be able to spin up a digital replica of you and query it until they find your vulnerability 😭.There are much worse outcomes, but I’ll just let your imagination wander. 🤔 💎 To avert this outcome, you’ll just have to keep growing in wisdom, starting with understanding what data is collected, how it can be used and most importantly how you can avoid having your data collected in the first place.✅ So you can make informed decisions.

4

An interesting and unintentional benefit of contemporary AI is that, because it is trained using our own everyday communications and data, it holds a mirror up to show us our own best and worst traits. The bigotry and sexism that shows up in AI's responses is there because that's what it saw. We can tweak parameters to remove the most overt cases, but as always, it's the subsurface discrimination that is most damaging."If you are letting AI learn from decisions that existing managers have historically made, and if those decisions have historically favored some people and disfavored others, Then that's what the technology will learn."https://lnkd.in/gQvNYGDV