Mindstream
Posts
What happened at the Seoul AI Summit?

What happened at the Seoul AI Summit?

+ AI can solve loneliness

Adam & Matt
May 30, 2024

PRESENTED BY

The fastest object ever made by humans was a manhole cover, covering a nuclear bomb test. It flew off at an estimated 125,000 mph - 5x the velocity needed to escape the Earth’s atmosphere.

What’s in store:

AI Summit for data usage regulation happened.
AI can give you a shoulder to cry on.
Around the web: T-Mobile to acquire key assets from US Cellular in a big move to boost rural 5G coverage, pending regulatory approval.
Learn why Skill Leap is the Netflix of AI.

Read time: 5 minutes

AI REGULATION

Seoul AI Summit - what happened?

The Bletchley Park AI summit in 2023 marked a significant step in AI regulation. Initially met with some bafflement, it soon became clear that AI regulation was in fact, worth a serious discussion.

The follow-up summit took place at a research park in Seoul last week. The challenge? Moving from talk to actual AI regulation.

At the end of the Seoul summit, the UK announced the creation of a global network of AI safety institutes.

These institutes, covering regions like the US, Canada, EU, and Asia, will share information on AI models, harms, and safety incidents.

This initiative builds on the success of the British institutes formed after the Bletchley Summit.

The UK Safety Institute shared results from tests on various models, including those from Anthropic.

These tests revealed the impact of their own jailbreaking attacks to test model safety.

Jack Clark from Anthropic highlighted that these institutes could embarrass companies into compliance by publicly revealing AI model flaws, a practice seen in academia.

Despite these advances, the institutes' power is limited to observation and reporting, which may not be enough to prevent AI harm.

The key takeaways:

Creating a global network of AI safety institutes is a major step forward.
These institutes will share information on AI models, harms, and safety incidents.
They aim to build on the success of British safety institutes formed after the Bletchley summit.
Public accountability will drive companies to address revealed flaws in AI models.

What does the future hold?

The EU and US institutes have set thresholds to define their oversight.

These thresholds create a problem for innovation, with companies striving to stay just below the limit to avoid regulatory scrutiny.

IBM's Christina Montgomery suggested that the focus will shift from model size to other factors like the number of users exposed to the AI.

A divide emerged on whether to regulate AI itself or its applications.

Some argue for regulating AI applications, comparing it to regulating electric motors rather than the motors themselves.

Concerns about superintelligent AI leading to catastrophic risks were raised, with some advocating for this to remain a central focus.

AI summits are so 2023.

Do you think global AI safety institutes will lead to meaningful change?

Vote for live results and see results + opinions from yesterday at the bottom of the email.

MASTER AI

Top Courses on ChatGPT and 50+ AI Tools

Ready to transform your work with AI?

Utilise the world's most comprehensive AI course library. Learn ChatGPT, Prompt Engineering, Midjourney, and 50+ essential tools tailored for entrepreneurs, marketers and creators.

What you'll get:

20+ Premium Courses
500+ Training Videos
50+ Downloadable Guides
5500+ Prompts
Private Community Group
New Content Added Weekly

Start here and see immediate results in your projects! →

On This Day

May 30th 1966 - NASA launches Surveyor 1, its first spacecraft to land on the Moon, marking a significant achievement in the American space program.

GPT of the Week

Photo Realistic GPT

Photo Realistic GPT generates highly detailed and vivid images based on user prompts. By enhancing these prompts with rich, imaginative details, it creates artistic and styled images.

Try it out →

AI PSYCHOLOGY

AI companionship, a solution for loneliness?

AI might be key to tackling loneliness, says Professor Tony Prescott from the University of Sheffield.

In his new book, The Psychology of Artificial Intelligence, he suggests that AI companions could be used as a form of social interaction, helping people break the cycle of loneliness and improve their social skills.

Loneliness poses serious health risks, including a 26% higher chance of premature death, and links to heart disease, dementia, stroke, depression, and anxiety.

In the UK, 3.8 million people suffer from chronic loneliness, while a Harvard study found that 36% of US adults and 61% of young adults feel seriously lonely.

Here’s what you should know:

AI Companionship: AI can provide social interaction to combat loneliness.
Health Impact: Loneliness increases the risk of premature death by 26% and is linked to various health issues.
Regulation Needed: Risks include users becoming overly dependent on AI, requiring regulation.

Can AI fill the human connection gap?

Prescott believes AI companionship could support people by providing personalised, stimulating interactions that boost self-worth and maintain or improve social skills.

However, he warns of potential risks, such as excessive dependence on AI, suggesting that regulation may be necessary.

As an expert in cognitive robotics, Prescott explores the link between the human brain and AI.

His research aims to recreate human perception, memory, and emotion in AI, contributing to our understanding of both human and artificial intelligence.

Prescott concludes that the ongoing partnership between psychology and AI will unlock further insights into both natural and artificial intelligence, helping to answer key questions about human existence and our future alongside AI.

I just need someone to acknowledge the 500 TikTok videos I send them.

AI TOOLS

MaxAI.me - Outsmart Most People with 1-Click AI

Discover MaxAI.me, one of the top 50 GenAI apps of 2024!

Best features:

Chat with the latest AI like GPT-4, Claude 3, and Gemini 1.5, all in one place.
Perfect your writing anywhere with 1-click AI without copy-pasting.
Save 90% of your reading & watching time with AI summaries.
Reply 10x faster with AI on email, social media, and messaging web apps.
Rapidly turn your visions into stunning images with AI art generators.

Join 1M+ users on MaxAI.me to get more done in less time!

Mindstream Picks

Apple's stock nears positive territory for the year, up 0.9% after a reported 52% surge in iPhone shipments in China for April.

A Chinese company associated with rocket maker Landspace, Shanghai Lanjian Hongqing Technology, has filed with the ITU to launch a satellite constellation of 10,000 units, aiming to expand its presence in the global satellite market.

After the closure of Liverpool's first dedicated gaming venue, leveltap, its gaming equipment including arcade machines, PCs, and monitors will be auctioned off online, ending Wednesday.

The Celestia blockchain's native token, TIA, surged 13% in the past 24 hours, trading at $10.30 with a market cap of $1.92 billion, ranking it as the 57th-largest cryptocurrency.

Don’t Miss - T-Mobile is set to acquire parts of US Cellular in a $4.4 billion deal, gaining wireless customers, stores, and some spectrum assets, while US Cellular retains most of its spectrum and all towers, leasing tower space to T-Mobile. This acquisition, expected to close by mid-2025, aims to enhance T-Mobile’s 5G coverage in rural areas and consolidate market competition, pending regulatory approval.