Key AI companies sign a White House ethics pledge to make a AI safer.

PLUS: How to up your game at problem formulation to improve your skills at prompt engineering

 

Today we cover the political news surrounding the big AI companies pledge to the Biden Administation as well as the way that businesses should be thinking about incorporating open-source software into their AI strategy.

We also highlight cool new tools focussed on text to image generation and short form video extraction. For this week’s take away, we cover how to get better at problem formulation as a skill to improve prompt engineering.

Let’s get to it!

🥽 3 TRENDS

Key players in the AI ecosystem sign White House pledge to make AI safer. (link)

US technology companies, including Google, OpenAI, Amazon, Anthropic, Inflection AI, Meta, and Microsoft, have made a public commitment to promote safety and transparency in the development of artificial intelligence.

This initiative, announced at the White House, involves voluntary commitments towards secure and transparent AI technology development, sharing more information about risk mitigation strategies, and increased security testing.

The White House described these commitments as a "critical step" towards responsible AI, while preparing an executive order and encouraging Congress to pass related legislation.

How businesses are dealing with the growing Open Source community and why they matter. (link)

The definition of "open source" in the realm of AI models may need to evolve, as models such as LLaMA2 are not entirely open due to certain restrictions, such as limitations on commercial use and restrictions against using model output to train another large language model.

The open source movement has witnessed several shifts, including a bifurcation caused by tension between commercial open source companies and cloud hyperscalers, leading to the creation of "Server-Side Public License" (SSPL) by companies like Elastic and MongoDB to prevent their products from being re-hosted as cloud services by companies like AWS.

Business owners will need to get smarter about open source software platforms and explore issues around promoting innovation, fostering transparency, and reducing costs by allowing access to a diverse community of developers who can collaboratively improve and tailor software to meet specific business needs.

Rumors begin to surface about Apple’s AI offering and how it is different from OpenAI and Google. (link)

Apple is developing artificial intelligence tools to compete with Google, OpenAI, and other major tech companies, and has built a chatbot internally referred to as "Apple GPT" according to Bloomberg's report.

This technology was built using a proprietary framework, codenamed "Ajax", which allows for the creation of large language models similar to OpenAI's ChatGPT and Google's Bard.

The deployment of this chatbot was temporarily halted due to security concerns about generative AI, but it has now been made accessible to more Apple employees for product prototyping.

Although Apple's chatbot doesn't currently offer features that distinguish it from other commercially available AI tools, the investment in this technology emphasizes the importance of artificial intelligence as a key strategic focus area for businesses to stay competitive and innovative in the technology sector.

🛠️ 2 TOOLS

 FOR THE CREATIVE
Adobe is working to make text to image content creation commercially safe. (link)

Adobe has developed a new AI software called Adobe Firefly, which is a robust text-to-image generator, that can quickly create marketing material from thoughts and ideas.

Unlike many AI art generators such as Midjourney, Stable Diffusion, and DALL-E2, which have been criticized for 'scraping' the internet for reference images regardless of copyright, Firefly has been trained on open-source images and content from Adobe Stock.

Adobe asserts that this approach will make Firefly commercially safe. The company plans to continually enhance Adobe Firefly by incorporating more features and resources from Adobe and other sources.

This development matters to businesses and creatives as it paves the way for the legal and ethical use of AI in creating artwork, enabling the production of unique designs without copyright infringement concerns, and enhancing creativity through AI-based tools.


 FOR THE MARKETER
Extract meaningful content from long-form video to use in social media posting. (link)

Munch is an all-in-one, AI-driven platform that allows users to extract impactful clips from long-form videos for repurposing across various social media platforms.

By leveraging advanced AI capabilities like GPT, OCR, and NLP, Munch analyzes video content in line with social and marketing trends, thereby generating suitable social posts for TikTok, Instagram, Twitter, LinkedIn, and YouTube Shorts.

The platform is designed to accommodate different user profiles, including brands, social media managers, media agencies, digital marketers, and content creators, providing them with valuable, actionable data for informed content decisions.

Not only does Munch assist in extracting key parts from videos, but it also helps in optimizing the video content for each platform, ensuring the action stays central, thereby simplifying the video editing and content creation process.

💡 1 TAKE AWAY

Getting good at problem formulation as the basis for prompt engineering.

Prompt engineering in AI has been garnering considerable discussion recently. Essentially, it is a mode of interaction between humans, acting as operators, and AI, functioning as assistants.

As an operator, I delegate a task to the AI and it responds with an outcome, which I can then fine-tune to align more closely with my desired result.

The ultimate objective in this iterative communication process is to shape the AI's output to mirror the concept I originally envisioned in my mind.

The ability to identify, analyze, and define problems, also known as problem formulation, is crucial in AI.

Rather than concentrating on specific words or phrases, we should aim to comprehend the "problem" we're attempting to "solve".

Break into small steps

Once the core problem is established, it needs to be broken down into smaller, actionable steps, which improves the AI's understanding and execution.

However, communication and language processing can be challenging with AI, sometimes requiring the use of analogies or proactive questioning to clarify tasks.

As AI’s language comprehension and processing improve, communication will become easier and the effectiveness of the output will depend on how well the tasks needed to solve problems are explained.

Thus, in the AI realm, our roles transform into curators, operators, and creative directors, with the responsibility to guide our AI assistants as efficiently as possible.

Thats a wrap!

We’ll see you again next week. Please send us your thoughts and any ideas you have to improve this content. If you have any questions you can reach out to us at [email protected]

Cheers,

The Simply Augmented Team