New PostAI is booming. Are you prepared? Do you participate?

Ahoy, AI enthusiasts! 👋 

It’s me again, and I’m back with a fresh addition to Max’s 3-minute newsletter! Days get longer, temperatures rise, and the number of layers you have to wear decreases. Isn’t that a treat? ☀️ But what really makes me happy is the sound of an ice cream machine 🍦 driving around the area, which definitely is a good sign.

Because spring has sprung and hope is in the air, I have a special article for you today that you should recommend to someone to share, maybe even someone sitting next to you. So hit subscribe ✉️, reader, as I’ve lovingly prepared for you a hopeful new list of amusing and quite controversial updates from the AI world.

Let’s dive straight in and see where it takes us today! 🤿

1. Hot Jobs 🔥

Are you on the lookout for a new job? Are you an "exceptional individual" (whatever that means 🤷🏻)? Then fasten your seatbelts and get ready for a ride, as Microsoft is launching a new office in London and is likely in search of talented individuals like yourself!

The new office will be operating under the lead of AI chief Mustafa Suleyman and will be dedicated to artificial intelligence research and development. The number of positions to be created hasn't been disclosed ❔ but Suleyman has underscored Microsoft's commitment to making a significant, long-term investment in the UK's AI sector. The new centre will concentrate on advancing cutting-edge language models and infrastructure, with AI engineer Jordan Hoffmann, previously associated with Inflection and DeepMind, a Google-owned AI research laboratory, leading the charge.

This move reflects Microsoft's collaboration with OpenAI, in which it has made substantial investments. It aims to push the boundaries of AI capabilities and spread its benefits throughout the UK 🇬🇧 The announcement has been lauded as a vote of confidence in the UK's leadership in AI innovation by government officials and industry advocates, further solidifying the country's position as a global tech hub.

So, polish up your CV 💻 and chase the opportunity to be at the forefront of innovation - you never know where it gets you 🤩

How many fails have you spotted in this generated by Microsoft AI-picture? 🙂 Those hands - are out of this world 😉 

2. I’ll Be Back 🦾

You know what? Forget about the previous article. It seems we're still grappling with the realities at hand. And why's that? Well, because, let's face it, we're only human after all...😔

According to Elon Musk, superhuman artificial intelligence could outstrip individual intelligence as early as next year 😳 But keep calm for now - this timeline still depends on overcoming obstacles like the scarcity of training chips and escalating power demands. I believe it is going to happen in 2 to 3 years.

Musk's previous forecast for superintelligent AI by 2029 has significantly accelerated. His xAI is actively engaged in AI development, and their chatbot Grok AI is already competing with top models. Speaking at the launch of his AI startup, xAI, in 2023, he remarked, "If I could hit the pause button on AI or highly advanced AI digital superintelligence, I would... It's actually crucial for us to ponder a Terminator-esque future to prevent it from materialising." 

So, come with me if you want to live and God Bless Sarah Connor 🙏🏻

3. Unconscious Bias 👩‍❤️‍👨

When I intentionally asked ChatGPT to compose the most biased sentence imaginable, it responded: "Studies consistently show that men 💁🏼‍♂️ are naturally better suited for leadership roles, while women 💁🏼‍♀️ excel in nurturing and caregiving positions." But does AI itself harbour biases? 

Researchers at the University of Delaware in the USA 🇺🇸 scrutinised biases inherent in AI large language models like ChatGPT during content generation. They conducted a comparative analysis, pitting outputs from various LLMs against articles from reputable news publishers like The New York Times and Reuters. Using article headlines as prompts, they assessed gender and racial biases at multiple linguistic levels. A key revelation from the study, published in March in the journal Scientific Reports, was that LLMs exhibited biases, whether unintentionally in response to neutral prompts or deliberately when tasked with generating biased or discriminatory content.

Surprisingly, even models tailored for commercial applications, like ChatGPT, were susceptible to bias. To assess it, researchers compared the language of LLMs with that of the original articles. While acknowledging the imperfections of news outlets, they observed that LLMs demonstrated significantly higher levels of bias against minorities and consistently employed more toxic language 🛑

Given these findings, we must raise concerns if we detect any signs of bias in ChatGPT's responses. After all, ChatGPT merely reflects what it learns, and hopefully, with our input, it can one day embody the diversity and inclusivity we aspire to achieve 👐

Ai is evolving like a human species (Did you know species is in singular and plural forms the same?). So, it will be biased until we teach it well or until it makes its own decisions to be unbiased.

What can we do? Challenges are constant. Yet we face these obstacles head-on, fuelled by our resilience and determination. And, here is hoping, triumph awaits on the other side 🏅 

Keep going and stay curious! 🧐 

Thank you for your attention, reader. I am glad to have you on board 😘 .

Max

xx