This story originally appeared in The Algorithm, a weekly newsletter about AI. To get news like this in your inbox first: Sign up here.
smart.
Is there anyone there?
AI telling common jokes. Researchers at Google DeepMind asked 20 professional comedians to write jokes and comedy performances using popular AI language models. Results were mixed.
Comedians said the tool was useful for creating an initial “vomit draft” that they could repeat and helped them structure routines. But AI hasn't been able to produce anything original, provocative, or decidedly entertaining. My colleague Rhiannon Williams has the full story.
As Tuhin Chakrabarty, a computer science researcher at Columbia University who specializes in AI and creativity, told Rhiannon, humor often relies on the surprising and incongruous. While creative writing requires writers to deviate from the norm, an LLM can only imitate this.
And this is becoming very evident in the way artists approach AI today. I have just returned from Hamburg, where I hosted one of the biggest creative events in Europe. The message I received from people I spoke to was that AI is too flawed and unreliable to completely replace humans and is best used as AI instead. Tools to enhance human creativity.
We are at a moment now where we are deciding how much creativity we are willing to give to AI companies and tools. Since the boom began in 2022 when DALL-E 2 and Stable Diffusion first appeared, many artists have raised concerns that AI companies are scraping their works without consent or compensation. Tech companies argue that everything on the public internet constitutes fair use, a legal doctrine that allows the reuse of copyright-protected material under certain circumstances. Artists, writers, imaging companies, and the New York Times have filed lawsuits against these companies, and it will likely be years before we get a clear answer as to who is right.
Meanwhile, the court of public opinion has undergone many changes over the past two years. Artists we interviewed recently said they were harassed and ridiculed for protesting the AI company's data scraping practices two years ago. The general public now knows more about the harms associated with AI. In just two years, the public has gone from being fascinated by AI-generated images to sharing viral social media posts about how to opt out of AI scraping. Until recently, this was an unfamiliar concept to most people. Businesses have also benefited from these changes. Adobe has been successful in promoting its AI products as an “ethical” way to use the technology without having to worry about copyright infringement.
There are also grassroots efforts to shift power structures in AI and give artists more power over their data. I've written about Nightshade, a tool created by researchers at the University of Chicago. This tool allows users to add invisible poison attacks to images, breaking the AI model when scraping them. The same team is developing Glaze, a tool that allows artists to mask their personal style from AI copycats. Glaze has been integrated into Cara, a new art portfolio site and social media platform that has seen a surge in interest from artists. Cara positions itself as a platform for people-created art. Filter AI-generated content. We gained almost a million new users in just a few days.
All of this is reassuring news for creative people who worry about losing their jobs to computer programs. The DeepMind study is a great example of how AI can actually help creators. It may take on some of the boring, mundane, and formal aspects of the creative process, but it can't replace the magic and ingenuity that humans bring. AI models are limited by their training data and forever reflect only the zeitgeist of the moment they were trained. You're getting old pretty quickly.
Now read the rest of the algorithm.
deep learning
Apple promises personalized AI in its private cloud. Here's how it works:
Last week, Apple unveiled its vision to enhance its product lineup with artificial intelligence. A key feature that will run across nearly all product lines is Apple Intelligence, a suite of AI-powered features that deliver personalized AI services while keeping sensitive data safe.
Why this is important:Apple says its privacy-focused system will first attempt to perform AI tasks locally on the device itself. When data is exchanged with a cloud service, it is encrypted and then deleted. This claim stands in implicit contrast to Alphabet, Amazon, Meta, etc., which collect and store enormous amounts of personal data. Read more from James O'Donnell here.
bits and bytes
How to Opt Out of Meta’s AI Training
If you post or interact with your chatbot on Facebook, Instagram, Threads, or WhatsApp, Meta can use your data to train generative AI models. Even if you don't use Meta's platform, data such as your photos may be scraped when others post them. Here's a quick guide on how to opt out: (MIT Technology Review)
Microsoft's Satya Nadella is building an AI empire
Nadella is going all in on AI. His $13 billion investment in OpenAI was just the beginning. Microsoft has become “the world’s most active accumulator of AI talent, tools, and technologies” and has begun building an in-house OpenAI competitor. (Wall Street Journal)
OpenAI has hired an army of lobbyists
As countries around the world consider AI legislation, OpenAI has begun hiring lobbyists to protect its interests. The AI company expanded its global affairs team from three lobbyists to 35 in early 2023 and plans to increase that to up to 50 by the end of the year. (Financial Times)
UK launches Amazon-powered emotion recognition AI cameras on trains
People traveling through Britain's largest train station are likely to have had their faces scanned by Amazon software without realizing it during an AI trial. London stations including Euston and Waterloo have tested AI-powered CCTV cameras to reduce crime and detect people's emotions. Emotion recognition technology is a very controversial technology. Experts say this is unreliable and simply doesn't work.
(mad)
Clearview AI uses your face. Now you can get shares in the company.
A facial recognition company that has been criticized for scraping images of people's faces from the web and social media without permission has agreed to an unusual settlement in a class action lawsuit against it. Instead of paying cash, it's giving Americans whose faces are in the data set a 23% stake in the company. (New York Times)
Elephants call each other by name
So cool! Researchers used AI to analyze the calls of two herds of African savannah elephants in Kenya. They discovered that elephants use specific vocalizations for each individual and recognize when other elephants are talking to them. (tutelar)