Posts in: AI

The Hidden Productivity Dilemma: AI Use in the Workplace 🤖💼

Using #AI tools at work can boost productivity ⚡, but it comes with social costs 😬. A study from Duke University, involving over 4,000 participants, reveals that colleagues often perceive AI users as lazy 🦥 and incompetent 🤷.

The crucial factor? The AI affinity of supervisors 👔.

In the study, fictional job applicants with AI skills were preferred only if the hiring manager also used AI tools – otherwise, they faced rejection 🚫.

„Evaluators who do not use AI regularly (less than weekly) evaluate candidates who do use AI as lazy.“

This highlights a critical point: Management plays a pivotal role in AI adoption within companies. Their own familiarity and comfort with AI tools can significantly influence how AI is perceived and integrated into the workplace. 🔍💡

Feeling a bit doomsday-ish? Then this post is for you!

#AI is advancing at warp speed, and researchers are still grappling with how these models really work. 🤖💨 Research revealed some creative strategies of these models (see my previous post), but what if constant training leads to hidden capabilities that suddenly erupt? What if the model goes rogue? 😱

Researchers warn of an intelligence explosion, power accumulation, disruption of democratic institutions, and more. Scary stuff, right?

But wait! They’ve also proposed a number of measures to address these challenges, but these are based on the sense of responsibility of the companies behind these large-scale models. So hopefully the paper will get some attention there. 🤞

Full research paper

Ever noticed how we struggle with rapid, non-linear change? 🌪️ Fasten your seatbelts, because groundbreaking AI-based developments could make your head spin!

Enter superhuman AI 🤖. Experts predict its impact over the next decade could outstrip the Industrial Revolution. Yeah, you heard right.

The team at AI Futures Project has crafted scenarios backed by trend extrapolations, wargames, expert feedback, and their previous forecasting wins. And the best part? It’s presented in a super intuitive way.

Don’t expect a crystal ball prediction. Think of it as a turbo boost into very near futures!

Great resource to dive in and get those mental gears turning! 💡

A more sovereign AI software setup

I I am experimenting with a "sovereign" software setup that leverages AI technology while strengthening data sovereignty and privacy. This setup combines the power of local frontend tools with advanced AI models hosted in Europe.

Key Components:

1. Local Frontend Tool: I am using AnythingLLM, a versatile tool that allows me to interact with AI models locally. This ensures that key data remains on my device, enhancing privacy and control.

2. OpenAI Compatible API: The frontend tool interfaces seamlessly with an OpenAI-compatible API, making it easy to integrate with various AI services.

3. IONOS AI Model Hub: For the backend, I leveraged the recently introduced IONOS AI Model Hub, hosted in Germany. This ensures that the data I transmit to the model remains within European jurisdiction, complying with stringent data protection laws.

4. European LLM: At the heart of the stack is OpenGPT-X / Teuken-7B, a European, open, multilingual large language model developed by Fraunhofer IAIS. Its a 7B class model, so has seven billion parameters and was trained from scratch with the 24 official languages of the EU.

Why This Setup Matters:

• Data Sovereignty: By keeping data local and within European servers, such a setup ensures compliance with GDPR and other data protection regulations.
• Performance: The combination of a local frontend and a robust backend has the potential to deliver high performance and low latency.
• Innovation: Using European-developed AI models supports local innovation and reduces dependency on foreign technologies.

Future Potential:

This setup is an experiment. The model is not on par with the latest models of global competitors, but as AI continues to evolve, maintaining technological sovereignty will become increasingly important.

By using local tools and European AI models, we can build resilient and compliant AI solutions. There must be sufficient resources for AI research and development in Europe to have a realistic chance of catching up with the best models!

But not every use case needs a huge LLM, so it's also important to understand how this setup works in the real world.

#AI #DataSovereignty #Innovation #EuropeanTech #DigitalTransformation #TechTrends

#FederatedLearning #FL, a decentralised #MachineLearning approach, enables banks to gain insights into customer behaviour without centralising sensitive data. This technology, while challenging to implement, offers personalised services, improved fraud detection, and accurate risk assessment.

While this is good for banks, it also opens interesting perspectives also in #InternationalCooperations handling of data! #FutureOfCooperation

See an example from finance here.

Thanks to Khang Vu Tien for pointing this out!

🔥 #OpenAI is going #OpenSource (sort of) - again. Sam Altman says a major “open weight” model is dropping this summer. Is this the beginning of the end for closed #AI? 👀

Read the full story @ Wired

The weird way an #LLM does the maths - pretty smart-ish!

Never has a technology been so widely adopted yet so mysterious in its inner workings. #Anthropic had to put some serious effort into a method, that gives a glimpse into how their model #Claude „thinks“.

So Anthropic’s deep dive reveals that while the model ends up with the correct answer, its internal process is anything but conventional. It uses quirky, approximate methods that differ starkly from the neat explanation it later provides.

When asked to add 36 and 59, Claude first approximates by summing “40ish and 60ish” and “57ish and 36ish” to arrive at a rough 92ish, then separately focuses on the last digits (6 and 9) to ensure the final number ends in 5. Combining these steps, it correctly computes 95. Pretty smartish. 😅

When asked how it arrived at this result, it gives you a common approach that you can find everywhere online, rather than what it actually did.

So, there’s another reason why it’s best to be careful and check what they say, rather than just believing them.

Full article in technology review

Even more detailed information on more examples in the research paper

This playbook is an inspiring resource with insights into creating impactful #DataScience and #AI #4good projects. It captures DataKind’s strategies and offers valuable lessons to improve the journey. Claiming to be continually updated with new insights and user feedback, it has the potential to become a go-to guide for those diving into #Data4Good initiatives.

🚀 Meet Europe’s new AI Factories: 🇦🇹 AI:AT | 🇧🇬 BRAIN++ | 🇫🇷 AI2F | 🇩🇪 JAIF | 🇵🇱 PIAST | 🇸🇮 SLAIF

With an investment of €485 million, these centres are providing SMEs and start-ups with access to advanced supercomputing, boosting Europe’s AI capacity.

AI Gigafactories, coming soon, will expand Europe’s AI power. Highly relevant! 🌍✨

See full press release.

🚀 Exciting times in #AI development! Showing the potential of #OpenSource! While #DeepSeek as such is open source, it is not yet possible to fully reproduce the model. Now, the #OpenR1 project on #HuggingFace is on a mission to fill the gaps in the DeepSeek-R1 pipeline and make this advanced AI approach accessible to all, allowing anyone to reproduce and build upon it. Work in progress, worth keeping an eye on! 🌐🤝

🔗 Check out the project here