- AI Odyssey
- Posts
- “AI” named word of the year
“AI” named word of the year
...Labeling won't solve AI's problems + Scarlett Johansson takes legal action against AI app...
Lavena Xu-Johnson
November 02, 2023
Hello. “AI” named word of the year. Purists might argue that an initialism formed by the words “artificial” and “intelligence” should be disqualified for Collins’ most notable word of 2023. But if nothing else, it confirms that society is at the peak of inflated expectations in the hype cycle before the eventual plunge into the trough of disillusionment.
— Lavena Xu-Johnson
⏰ Top 5 Headlines
🔗Link: Scarlett Johansson takes legal action against AI app that ripped off her likeness in advertisement.
🔗Link: LinkedIn passes 1 billion members and launches new AI chatbot to help you get a job.
🔗Link: IAC warns regulators generative AI could wreck the web.
🔗Link: Microsoft’s new AI-powered Office assistant is here — if you’ve got $9,000 and 300 friends.
🔗Link: Stability AI, Tiktok, and several governments signed an agreement to fight AI-generated CSAM.
Presented by Brilliant
Future-proof yourself by mastering the concepts behind AI, neural networks, data science, and more on Brilliant.
Brilliant makes it easy to level up—with bite-sized, interactive lessons on cutting-edge topics that help build your knowledge and problem-solving skills in minutes a day.
Join over 10 million professionals and lifelong learners and start your 30-day free trial today. Plus, AI Creator readers can get a special 20% off a premium annual membership.
💬 Deeper Dives
‘It’s not clear we can control it’
Bletchley Park AI summit raises concerns about the catastrophic risk of AI technology to humanity, with the UK, US, EU, and China making a landmark declaration.
Elon Musk, Tesla CEO, and co-founder of OpenAI, labels AI as a threat to humanity and discusses the need to guide it in a beneficial direction.
Mustafa Suleyman, co-founder of DeepMind, suggests a possible pause in AI development over the next five years but believes current models like ChatGPT do not pose a significant threat.
King Charles emphasizes the importance of addressing AI challenges, drawing parallels with climate change efforts and the need for global collaboration.
UK Technology Secretary Michelle Donelan discusses AI's potential to reduce tedious administrative tasks and calls for adapting education and skills to AI-related job changes.
Věra Jourová, the European Commission's vice-president, highlights differences in AI regulation approaches between the UK, US, and EU, with the UK focusing on potential risks before regulation.
Rishi Sunak, UK Chancellor of the Exchequer, rules out immediate AI legislation, emphasizing the need for a better understanding of the technology.
Matt Clifford, organizer of the summit, demonstrates how powerful AI models can be used for malicious purposes, such as launching phishing attacks, emphasizing the importance of addressing real-world AI capabilities.
Labeling won't solve AI's problems
As AI becomes more integrated into our digital tools and daily lives, labeling AI-generated content is becoming less clear and effective. The Biden administration and Congress are considering labeling requirements, but as AI becomes a collaborator, labeling loses its significance.
Collaboration blurs lines between human and AI: When people use AI as a collaborative tool, such as in programming or content creation, clear distinctions between human and AI work become inadequate. This creates questions about authorship and provenance.
Labeling issues extend to various media: The challenges of labeling extend beyond text and affect images, code, and even voice-acting in video games. As AI becomes an integral part of creative tools, the concept of "AI-generated" content loses its meaning, leading to ongoing uncertainty in classification.
(Read more: Axios)
The race is on to bring AI to weather forecasting
The global race to enhance weather forecasting is now centered on the integration of artificial intelligence, with the UK Met Office and NOAA in the US collaborating with AI experts to develop cost-effective and highly accurate weather forecast models.
AI-driven weather models, based on historical data and machine learning, offer the potential to overcome the limitations of supercomputer-based numerical models, significantly reducing computation time and expenses, though skeptics raise concerns about their susceptibility to errors and biases.
(Read more: Axios)
AI generated images are biased, showing the world through stereotypes
In an exclusive report, The Washington Post delves into the disconcerting world of AI image generators, shedding light on how these tools perpetuate harmful stereotypes. These AI image generators, such as Stable Diffusion and DALL-E, inadvertently amplify biases in gender and race, despite attempts to purify the data they rely on. The troubling results include representations of attractive people as predominantly young and light-skinned, while Muslim people are often depicted as men with head coverings. These generated images underscore the profound impact AI can have in perpetuating our society's worst stereotypes, as they reflect the biases inherent in the data that fuels these technologies.
(Read more: Washington Post)