Is Artificial Intelligence really bad? – Sam Altman’s perspective on the state of the art and the role of artificial intelligence in generating AI
The pace of change in generative AI right now is insane. Just four months ago, Openai released its flagship product to the public. It took two months to reach 100 million users. TikTok, the internet’s previous instant sensation, took nine. There are a lot of different bot clones and new plug-ins already on the market, and Bard is one of them. GPT-4, the new version of OpenAI’s model released last month, is both more accurate and “multimodal,” handling text, images, video, and audio all at once. The latest release of MidJourney has given us the sensation of Donald Trump being arrested and the Pope looking like a damsel in distress, which make it clear that you will soon have to treat every single image you see.
In order to truly create public benefit, we need mechanisms of accountability. The world needs a generative AI global governance body to solve these social, economic, and political disruptions beyond what any individual government is capable of, what any academic or civil society group can implement, or any corporation is willing or able to do. Companies and countries have already cooperated to hold themselves accountable for technological outcomes. Independent, well-funded expert groups can make decisions for the public good. An entity like this is tasked with thinking of benefits to humanity. generative AI is already facing some fundamental issues and we should build on these ideas to address them.
The closest person to a chief architect is the CEO of OpenAI, Sam Altman, who in an interview with The New York Times called the Manhattan Project the level of ambition we aspire to. The others are Tristan Harris and Aza Raskin of the Center for Humane Technology, who became somewhat famous for warning that social media was destroying democracy. They are now going around warning that generative AI could destroy nothing less than civilization itself, by putting tools of awesome and unpredictable power in the hands of just about anyone.
To be clear, neither Harris nor Raskin agree with Altman on the issue of Artificial Intelligence being able to destroy civilization. He is trying to ensure that the tools are developed with guardrails, because the technology is unstoppable, and he has no choice but to keep moving because of the unstoppable technology. It’s a mind-boggling mix of faith and fatalism.
I agree with the fact that the tech is unstoppable. At the moment, the guardrails being put in place are weak and laughably weak. It would be a fairly trivial matter, for example, for companies like OpenAI or MidJourney to embed hard-to-remove digital watermarks in all their AI-generated images to make deepfakes like the Pope pictures easier to detect. Artists are able to attach information to artificial intelligence-generated pictures if they use the Content Authenticity Initiative’s protocol. I don’t see any of the major generative Artificial Intelligence companies joining those efforts.
Every time you post a photo, respond on social media, make a website, or possibly even send an email, your data is scraped, stored, and used to train generative AI technology that can create text, audio, video, and images with just a few words. This has real consequences: OpenAI researchers studying the labor market impact of their language models estimated that approximately 80 percent of the US workforce could have at least 10 percent of their work tasks affected by the introduction of large language models (LLMs) like ChatGPT, while around 19 percent of workers may see at least half of their tasks impacted. The labor market is shifting immediately with image generation as well. It is possible that the data you created may be putting you out of a job.
The internet is a public resource and it is reasonable to say that technology should be open to all. Critics said that GPT-4 lacked any information or specifications that would allow someone outside of the organization to replicate, test, or verify the model. Some of the companies have received a lot of funding from other corporations. This is a sign that the companies will look for profits over the public’s benefit.
The Berkman Klein Center for Internet and Society at Harvard University has a member who works on Responsible Artificial Intelligence. She was the director of machine learning ethics at the time.
The public good isn’t likely to be ensured through code transparency alone. There is little conceivable immediate benefit to a journalist, policy analyst, or accountant (all “high exposure” professions according to the OpenAI study) if the data underpinning an LLM is available. Many laws, like the Digital Services Act, require some companies to give expert auditor access to their code and data. And open source code can sometimes enable malicious actors, allowing hackers to subvert safety precautions that companies are building in. Transparency is a laudable objective, but that alone won’t ensure that generative AI is used to better society.
After World War II there was a credible and significant fear of nuclear technologies going rogue. The widespread belief that society had to act collectively to avoid global disaster echoes many of the discussions today around generative AI models. The US and the United Nations lead countries around the world in forming the International Atomic Energy Agency, an independent body free of government and corporate affiliation that would provide the solutions to the far-reaching ramifications and seemingly infinite capabilities of nuclear technologies. Nuclear safety and security is one of the main areas that it operates in. It provided critical resources, education, testing, and impact reports after the disaster in 2011; it was helpful to ensure ongoing nuclear safety. The agency is limited because it relies on member states to comply with its standards and guidelines, and on their cooperation and assistance to carry out its mission.
AI Video Generators Are Nearing a Crucial Tipping Point: How Artificial Intelligence Has Been Learned in the First Months
The Oversight Board in Facebook is trying to balance transparency with accountability. The Board members are an interdisciplinary global group, and their judgments, such as overturning a decision made by Facebook to remove a post that depicted sexual harassment in India, are binding. This model isn’t perfect either; there are accusations of corporate capture, as the board is funded solely by Meta, can only hear cases that Facebook itself refers, and is limited to content takedowns, rather than addressing more systemic issues such as algorithms or moderation policies.
But you only need to look at how advanced images from Midjourney and Dream Studio are now to sense where AI video is heading—and how difficult it may become to distinguish real clips from fake ones. Of course, people can manipulate videos using existing technology, but it is not easy to pull off.
Craiyon was an open source knockoff of the then carefully restricted DALL-E 2 image generator from OpenAI, the company behind ChatGPT. The tool showed how artificial intelligence could turn a text into something like photos or illustrations. Since then, DALL-E has become open to everyone, and programs like Midjourney and Dream Studio have developed and honed similar tools, making it relatively trivial to craft complex and realistic images with a few taps on a keyboard.
As engineers have tweaked the algorithmic knobs and levers behind these image generators, added more training data, and paid for more GPU chips to run everything, these image-making tools have become incredibly good at faking reality. Alex Jones is at a gay pride parade and the Ark of the Covenant is at a yard sale, all pictured on a subreddit devoted to strange Artificial Intelligence images.
Widespread availability of this technology, combined with its sophistication, causes us to rethink how we view online imagery, as was demonstrated after an artificial intelligence-made image showing Donald Trump’s arrest went viral. The incident led Midjourney to announce that it would no longer offer a free trial of its service—a fix that might deter some cheapskate bad actors but leaves the broader problem untouched.
Source: https://www.wired.com/story/ai-video-generators-are-nearing-a-crucial-tipping-point/
How to make a video from an old one without changing it: a new technique for applying stylistic changes to a landscape using artificial intelligence and the Luma app
Runway ML, a startup that’s developing AI tools for professional image and video creation and editing, this week launched a new more efficient technique for applying stylistic changes to videos. I used it to create this video of my cat walking through a cloudscape, from an existing video.
Different machine learning techniques open new possibilities. A company called Luma Artificial intelligence uses neural radiance fields to turn 2D photographs into detailed 3D scenes. You can play with a fully interactive 3D scene when you feed a few snapshots into the company’s app.
For now, the instinct to trust video clips is mostly reliable, but it might not be long before the footage we see is less solid and truthful than it once was.