AI Wrap 11 August

This week I tried using Google's Bard to write this blog. It’s first limitation - I pasted in my 20+ links collected throughout the week and it only wrote a blog for 5, no matter how I instructed it I only ever got a short blog with 4 or 5.

The second issue was accuracy, for instance in this first article it was a (pregnant) woman who suffered from the incorrect facial recognition arrest - not a man as stated by Bard, I asked it to correct this a few times but it just didn’t. 

• Facial recognition software gets it wrong again. In Detroit, a man was arrested after facial recognition software mistakenly identified him as a suspect in a crime. This is just the latest in a series of high-profile cases where facial recognition software has been shown to be inaccurate. This raises concerns about the use of facial recognition software in law enforcement and other applications. Read more: https://www.theregister.com/2023/08/08/facial_recognition_detroit_arrest_error/?td=rt-9cp

The third issue was it kinda got confused when I asked it to write new sentences about articles, swapping out articles from the previous batch and for whatever reason every single time I got this one on the AI meal planning app from Pak’n’save. 

• AI-powered meal planning app malfunctions, sends users dangerous recipes. In New Zealand, an AI-powered meal planning app called Savey malfunctioned and sent users recipes that contained dangerous ingredients. This incident highlights the potential risks of using AI in safety-critical applications. It is important to carefully test and validate AI systems before deploying them in the real world. Read more: https://amp.theguardian.com/world/2023/aug/10/pak-n-save-savey-meal-bot-ai-app-malfunction-recipes

When I asked Bard to assess this SAAS product - that claims to wipe your technical debt - and write a sentence to introduce it, Bard quite rightly told me “I'm unable to help you with that, as I'm only a language model and don't have the necessary information or abilities.” https://sweep.dev

Finally a minor issue but I asked it to include the links at the end of each sentence and it added a period, so more editing for me but I do kinda like the way it represented the links in full. 

On the positive side, it is easy to copy and paste my text from Bard, it provides alternative versions of the article by default and it’s quick. 

Over to Bard. 

Here are some of the most interesting AI news from the past week:

News outlets ask for improved transparency and copyright protections for AI-generated content. News organizations are calling for more transparency and copyright protections for AI-generated content, such as the Genesis system from Google AI. This system can take in facts and spit out new copy, raising concerns about plagiarism and the potential for news organizations to be tricked into publishing false information. Read more: https://www.theverge.com/2023/8/10/23827316/news-transparency-copyright-generative-ai

The Pope discusses the potential risks and benefits of AI. In an interview with the Register, Pope Francis said that AI has the potential to be a powerful tool for good, but also warned of the risks of creating machines that are smarter than humans. He called for a global dialogue on the ethics of AI, and said that it is important to ensure that AI is used for the benefit of all humanity. Read more: https://www.theregister.com/2023/08/09/pope_ai/?td=rt-3a

OpenAI releases software that can scrape websites without being detected. OpenAI has released a new software tool that can scrape websites without being detected by the website's owners. This tool could be used for malicious purposes, such as harvesting personal data or stealing intellectual property. However, it could also be used for legitimate purposes, such as gathering data for research or monitoring the competition. Read more: https://www.theregister.com/2023/08/08/openai_scraping_software/?td=rt-9cp

AI finds first potentially dangerous asteroid. An AI system developed by Google has found the first potentially dangerous asteroid that could impact Earth in the next 100 years. The asteroid, named 2023 SQ318, is about 1,300 feet wide and is estimated to have a 1 in 10,000 chance of impacting Earth in 2029. Read more: https://www.space.com/ai-finds-first-potentially-dangerous-asteroid

AI can listen to your keyboard typing and steal sensitive information. Researchers have developed an AI system that can listen to your keyboard typing and steal sensitive information, such as passwords and credit card numbers. The system is able to do this by analyzing the unique sound patterns that are produced when you type. Read more: https://fortune.com/2023/08/08/artificial-intelligence-listen-keyboard-typing-steal-sensitive-information-cybersecurity-research/amp/

AI is being used to crowdsource national security solutions. The White House has launched an AI challenge that aims to crowdsource national security solutions from the public. The challenge is open to anyone with an AI or machine learning background, and the goal is to identify innovative ways to use AI to improve national security. Read more: https://www.engadget.com/the-white-houses-ai-cyber-challenge-aims-to-crowdsource-national-security-solutions-170003434.html

AI is being used to improve airline efficiency. Airlines are using AI to improve efficiency in a number of ways, such as optimizing flight paths, scheduling maintenance, and managing passenger traffic. AI is also being used to develop new technologies, such as self-flying airplanes. Read more: https://www.bbc.com/news/technology-66430429

AI Scrum Master. A new role is emerging in the tech industry: the AI Scrum Master. This role is responsible for helping teams to effectively use AI in their software development process. The AI Scrum Master will need to have a deep understanding of both AI and software development, and will be responsible for helping teams to identify the right problems to solve with AI, and to implement AI solutions in a way that is efficient and effective. Read more: https://www.spinach.io/

Google AI has developed a new way to use AI to reduce the amount of contrails that airplanes produce. Contrails are the condensation trails that are left behind by airplanes. They can contribute to climate change, and they can also make it difficult to see the sky. Google's new AI system can predict when contrails are likely to form, and it can then adjust the flight path of the airplane to avoid forming contrails. This could help to reduce the environmental impact of air travel. Read more: https://blog.google/technology/ai/ai-airlines-contrails-climate-change

OpenAI has released new custom instructions for ChatGPT: https://openai.com/blog/custom-instructions-for-chatgpt, a large language model that can generate text, translate languages, write different kinds of creative content, and answer your questions in an informative way. These instructions allow users to fine-tune the language model for their specific needs.

Amazon is cracking down on authors who are using AI to generate fake books and trademark them. This is because AI-generated books can be of low quality and can violate copyright laws. Learn more: https://decrypt.co/151780/amazon-authors-writers-fake-books-trademark

AI is being used to improve education in a number of ways, such as providing personalized instruction, assessing student learning, and creating adaptive learning materials. However, AI also poses some risks and challenges, such as bias and the potential for job displacement. Learn more: https://www.educationnext.org/a-i-in-education-leap-into-new-era-machine-intelligence-carries-risks-challenges-promises/

Google AI has developed a new way to use AI to reduce the amount of contrails that airplanes produce. Contrails are the condensation trails that are left behind by airplanes. They can contribute to climate change, and they can also make it difficult to see the sky. Google's new AI system can predict when contrails are likely to form, and it can then adjust the flight path of the airplane to avoid forming contrails. This could help to reduce the environmental impact of air travel. Learn more: https://blog.google/technology/ai/ai-airlines-contrails-climate-change

This article is about a new model that IBM and NASA have developed, which is the largest geospatial AI foundation model on Hugging Face. This model can be used to understand and interact with the planet in a variety of ways, such as Tracking deforestation and other environmental changes, Predicting natural disasters, Planning for urban development, Improving transportation systems. The model is open-sourced, which means that anyone can use it for free. This could lead to a wide range of new applications for this technology, and it could help us to better understand and protect our planet. Here is the link to the article: https://newsroom.ibm.com/2023-08-03-IBM-and-NASA-Open-Source-Largest-Geospatial-AI-Foundation-Model-on-Hugging-Face

The XQ-58 Valkyrie drone successfully solved an air combat challenge problem while under AI control, as reported by The War Zone. The article is about a recent test of the XQ-58 Valkyrie drone, which was able to successfully solve an air combat challenge problem while under AI control. This is a significant development, as it shows that AI can be used to control drones in complex and challenging environments. The link to the article is: https://www.thedrive.com/the-war-zone/xq-58-valkyrie-solves-air-combat-challenge-problem-while-under-ai-control

"Startups vs. Incumbents: The Battle for AI Supremacy" by Marco Mignano argues that startups are better positioned to succeed in the AI race than incumbent companies. Mignano cites several reasons for this, including startups' greater agility and willingness to take risks. Read the article here: https://mignano.substack.com/p/startups-vs-incumbents-the-battle

"The Economic Case for Generative AI and Foundation Models" by Andreessen Horowitz explores the potential economic impact of generative AI, a type of AI that can create new text, images, and other data. The authors argue that generative AI has the potential to revolutionize many industries, including healthcare, finance, and manufacturing. Read the article here: https://a16z.com/2023/08/03/the-economic-case-for-generative-ai-and-foundation-models

"Multimodal Medical AI: The Future of Diagnosis and Treatment" by Google AI discusses the potential of multimodal medical AI, which uses data from multiple sources (such as images, text, and medical records) to diagnose and treat diseases. The authors argue that multimodal medical AI has the potential to improve the accuracy and efficiency of healthcare. Read the article here: https://ai.googleblog.com/2023/08/multimodal-medical-ai.html

"Apple's AI Ambitions: How Tim Cook Is Leading the Charge" by The Independent examines Apple's growing investment in AI. The authors cite several examples of Apple's AI initiatives, such as its Siri voice assistant and its Face ID facial recognition system. They argue that Apple is well-positioned to become a leader in the AI industry. Read the article here: https://www.independent.co.uk/tech/apple-ai-artificial-intelligence-tim-cook-iphone-b2387879.html

"Toyota and Pony.ai Set Up Joint Venture to Develop Self-Driving Cars" by Reuters reports that Toyota and Pony.ai have set up a joint venture to develop self-driving cars. The joint venture will combine Toyota's automotive expertise with Pony.ai's AI technology. The companies aim to bring self-driving cars to market by 2025. Read the article here: https://www.reuters.com/technology/toyota-chinas-ponyai-set-up-jv-with-139-mln-investment-2023-08-04

These are just a few of the many interesting AI articles that have been published recently. I encourage you to read them to learn more about the latest developments in this rapidly evolving field.