philip lelyveld The world of entertainment technology

Brace Yourself for the 2024 Deepfake Election

No matter what happens with generative AI, its disruptive forces are already beginning to play a role in the fast-approaching US presidential race.

... Republicans recently used AI to generate an attack ad against Biden. The question is, what will happen when anyone can open their laptop and, with minimal effort, quickly create a convincing deepfake of a politician? 

There are plenty of ways to generate AI images from text, such as DALL-E, Midjourney, and Stable Diffusion. It’s easy to generate a clone of someone’s voice with an AI program like the one offered by ElevenLabs. Convincing deepfake videos are still difficult to produce, but Ajder says that might not be the case within a year or so.

“To create a really high-quality deepfake still requires a fair degree of expertise, as well as post-production expertise to touch up the output the AI generates,” Ajder says. “Video is really the next frontier in generative AI.” ...

That includes companies like Runway, Google, and Meta. Once one company releases a high-quality version of a text-to-video generative AI tool, we may see many others quickly release their own versions, as we did after ChatGPT was released. Farid says that nobody wants to get “left behind,” so these companies tend to just release what they have as soon as they can. ...

So what can be done about this problem? One solution is something called C2PA. This technology cryptographically signs any content created by a device, such as a phone or video camera, and documents who captured the image, where, and when. The cryptographic signature is then held on a centralized immutable ledger. This would allow people producing legitimate videos to show that they are, in fact, legitimate. ...

One last possibility for solving this problem would be to develop a sort of instant fact-checker for social media users. Aviv Ovadya, a researcher at the Berkman Klein Center for Internet & Society at Harvard, says you could highlight a piece of content in an app and send it to a contextualization engine that would inform you of its veracity. ... You could have it cross-referenced with sources you can trust.” ... The problem, he says, is that many founders he has spoken with simply don’t see a lot of money in developing such a tool.  ...

See the full story here: https://www.wired.com/story/chatgpt-generative-ai-deepfake-2024-us-presidential-election/

Joint meeting of Congress to be filmed in 360° virtual reality

House Speaker Kevin McCarthy (R-Calif.) will use a virtual-reality camera to record today’s joint meeting with South Korean President Yoon Suk-yeol, his office tells Axios. 

Why it matters: This is the first time a 360° camera has ever been used in the Hall of the House, providing unprecedented access to a joint meeting of Congress, the speaker's office says.

The GoPro MAX will be positioned on a stand, midway down the center aisle.

• McCarthy aides say the VR cut of the joint meeting will be released next week — with more congressional VR experiences to follow.

See the full story here: https://www.axios.com/2023/04/27/virtual-reality-congress-south-gopro

How augmented reality will resolve the prison guard shortage crisis

... AR to train prison guards for "what they will see, hear, smell, and experience once they enter the actual prisons." To be "ready for the assaults from the prisoners that have become more common. " ...

See the full story here: https://www.verdict.co.uk/prison-staffing-shortage-ar/https://www.verdict.co.uk/prison-staffing-shortage-ar/

Character AI Raises $150M at $1B Valuation for Conversational Generative AI Platform

... Character AI hosts a conversational generative AI hub with chatbots mimicking a wide variety of real and fictional personalities as famous or personal as the designer chooses. The LLM filtered through that personality can then help users with specific services or general assistance depending on their choices. ...

See the full story here: https://voicebot.ai/2023/04/26/character-ai-raises-150m-at-1b-valuation-for-conversational-generative-ai-platform/

OpenAI Previews ChatGPT Business, Unveils Privacy Controls

... ChatGPT data settings can be adjusted by logging in to your account and “selecting the three dots next to your email address in the bottom-left corner of the screen. From there, click Settings > Show and toggle off the Chat History & Training setting to turn off conversation history,” The Verge explains. You can also select the new Export data option using the three-dot menu.

“Once you’ve disabled your chat history, you’ll notice that none of your new conversations save to your history bar on the left side of the screen,” writes The Verge, adding that “you can quickly turn the option back on again by hitting the green Enable chat history button that appears in the column.” ...

See the full story here: https://www.etcentric.org/openai-previews-chatgpt-business-unveils-privacy-controls/

The Politics of Language Models

... Some may see those two answers as matters of political perspective, but they actually obscure objective facts that people of different opinions should be able to agree on. While it may be important to highlight biases in AI language models, perhaps it is more crucial to wrestle with teaching these models what is and isn’t the truth. ...

See the full story here: https://link.wired.com/view/5cc9e2393f92a477a0e9b299imuhc.154u/952e3409

WEIRD AI: Understanding what nations include in their artificial intelligence plans (Brookings Inst.)

... we warned that China is already in the lead on the achievement of its national AI goals and predicted that it would continue to surpass the U.S. and other countries. News has now come that China is planning on doubling its investment in AI by 2026, and that the majority of the investment will be in new hardware solutions. The U.S. State Departmentalso is now reporting that China leads the U.S. in 37 out of 44 key areas of AI.  In short, China has expanded its lead in most AI areas, while the U.S. is falling further and further behind. ...

In this blog, the countries that are part of our study include 34 nations that have produced public AI policies, as identified in our previous blog posts: Australia, Austria, Belgium, Canada, China, Czechia, Denmark, Estonia, Finland, France, Germany, India, Italy, Japan, South Korea, Lithuania, Luxembourg, Malta, Mexico, Netherlands, New Zealand, Norway, Poland, Portugal, Qatar, Russia, Serbia, Singapore, Spain, Sweden, UAE, UK, Uruguay, and USA. ...

For each, we examine six key elements in these national AI plans—data management, algorithmic management, AI governance, research and development (R&D) capacity development, education capacity development, and public service reform capacity development—as they provide insight into how individual countries approach AI deployment. In doing so, we examine commonalities between culturally similar nations which can lead to both higher and lower levels of investment in each area. ...

We do this by exploring similarities and differences through what is commonly referred to as the WEIRD framework, a typology of countries based on how Western, Educated, Industrialized, Rich, and Democratic they are. In 2010, the concept of WEIRD-ness originated with Joseph Henrich, a professor of human evolutionary biology at Harvard University.  ...

See the full article here: https://www.brookings.edu/blog/techtank/2023/04/25/weird-ai-understanding-what-nations-include-in-their-artificial-intelligence-plans/

Opinion The next level of AI is approaching. Our democracy isn’t ready.

... we need to strengthen the tools of democracy itself. A pause in further training of generative AI could give our democracy the chance both to govern technology and to experiment with using some of these new tools to improve governance. The Commerce Department recently solicited input on potential regulation for the new AI models; what if we used some of the tools the AI field is generating to make that public comment process even more robust and meaningful?

We need to govern these emerging technologies and also deploy them for next-generation governance. But thinking through the challenges of how to make sure these technologies are good for democracy requires time we haven’t yet had. And this is thinking even GPT-4 can’t do for us.

See the full editorial here: https://www.washingtonpost.com/opinions/2023/04/26/artificial-intelligence-democracy-danielle-allen/

Amnesty Int’l: EU: European Union must protect human rights in upcoming AI Act vote

Ahead of the European Parliament’s vote on the AI Act in May, the European Union (EU) has a significant opportunity to regulate artificial intelligence (AI) technologies in order to protect and promote human rights, said Amnesty International in an open letter to Members of Parliament’s leading committees. 

“The AI Act offers EU lawmakers an opportunity to put an end to the use of discriminatory and rights-violating artificial intelligence (AI) systems,” said Mher Hakobyan, Advocacy Advisor on Artificial Intelligence Regulation at Amnesty International. 

“The EU must ban the use of discriminatory AI systems which disproportionately affect people from marginalized communities, including migrants, refugees and asylum seekers. Such technologies profile people and communities, claiming to ‘predict’ crimes, or ‘identify’ people who supposedly pose a security risk, even leading to them being denied the right to asylum. EU lawmakers must not miss this opportunity to prohibit the use of certain AI-based practices and protect the rights of migrants, refugees, and asylum seekers against harmful aspects of AI. ...

“The AI Act should also address the development of European technologies that are exported to third countries. Firstly, AI systems that are prohibited in Europe should not be allowed to be exported abroad. Secondly, permitted high-risk technologies that are exported must meet the same regulatory requirements as high-risk technologies sold in the EU.”  ...

Key dates:

11 May: Planned committee vote on the AI Act 

See the full story here: https://www.amnesty.org/en/latest/news/2023/04/eu-european-union-must-protect-human-rights-in-upcoming-ai-act-vote/

Berkeley Talks: ChatGPT developer John Schulman on making AI more truthful

... part of the reason for generating untruthful content, which he calls “hallucinations,” is because the model doesn’t know that it’s allowed to say “I don’t know” or express uncertainty. If you tell a chatbot that it’s allowed to do that, he says, that partially fixes the problem. ...

“And then, another set of hallucinations, you could say, is that it’s just guessing wrong. ...

... reinforcement learning is part of the solution. ...

See the full story here: https://news.berkeley.edu/2023/04/24/berkeley-talks-chatgpt-developer-john-schulman/