top of page
  • Writer's picturelauralikespi

AI News - Fri 25th Aug 2023

It might be the last week of the summer holidays in the UK, with a bank holiday scheduled, but the AI news is not taking a rest. The focus this week has shifted to AI in schools (unsurprising as we brace for the new school year).

An representation of AI created by DeepMind


Companies to Watch


Poolside, a US based AI company creating a chatbot (similar to ChatGPT) to help write code have raised $126mil and are moving to Paris. Interestingly one of their founders is the previous CTO of Github (Read more)

The current website of

Portkey AI has raised $3mil to simplify the process of building with generative AI (Read more)


Interesting Reads

  • Wired's article on The Myth of Open Source AI.

  • The Conversation's piece on the current discussions around AI (great for those who want a bit more context and background on AI).

  • Corndel (the skills partner) has realised a report, Data Report 2023: Better Decisions, Realised. Ironically, you can access the report by giving them your data. It is worth a read, but there is a useful summary written by Silicon.

  • (For the LLM nerds out there) An academic paper on Giraffe: Adventures in Expanding Context Lengths in LLMs is worth a read.


Good to Know

Microsoft and Databricks Exploring OpenSource AI (Apparently)

In what could be a threat to both OpenAI and Meta, Microsoft (who have a large stake in OpenAI) have been discussing diversifying their AI offering with Databricks. This was originally reported in the Information last Friday, but this is behind a paywall.

Databricks help companies with all things data deployment (including creating custom machine learning and AI models). They already have a partnership with Microsoft and have recently acquired an OpenAI competitor MosaicML (Read more)

Google's DeepMind and Artists Visual AI

The creative team at DeepMind have been working with artists to create "diverse, accurate and accessible representations of AI" according to Jonathan Fildes (LinkedIn post below). This includes creating stock photos which can be used via Unsplash and other platforms (we have added some to this article) (Read more)

AI Safety Summit Scheduled

An AI Safety meeting has been arranged in November at Bletchley Park (home to the code breakers, including Alan Turing, during WW2) (Read more)

It is rumoured the UK and Japan have clashed over the UK's plans to invite China to this summit (Read more)

AI and Copyright Laws

"A US District Court has published its decision in Thaler v. The Register of Copyrights, reaffirming the longstanding principle that human authorship is a requisite for copyright protection" posted Barry Scannell (LinkedIn)

This law seems like a win for the Writers who are currently striking, as it means production studios will have incentive to have a human-writer involved in the process (Read more)

Meta Releases Code Llama

Meta has released an AI-code writing tool, built on top of their Open Source LLM Llama2, called Code Llama. Meta are claiming this is performing better than other LLM based coding tools (eg Github's co-pilot) (Read more)

Midjourney's New Feature

Midjourney (do check out the website) is a generative AI image tool (so a competitor to OpenAI's Dall-e). This week they announced inpainting which allows creators to edit the image when the AI is finished generating. People are hyped (especially on Twitter, so take what you will from that) (Read more)

South Korean Internet Giant Creates Own GenAI Tool

Naver (who are a search engine company) have announced their own Large Language Model HyperCLOVA X, chat CLOVA X and a generative AI search engine, Cue. According to Tech Crunch, these tailored AI solutions are aimed at Southeast Asia, the Middle East and other non-English speaking countries (Read more)

American Companies Are Obsessed With AI - Even If They Don't Understand Why

American companies have published their quarterly reports, and the Washington Post has analysed them - finding over 1,000 companies mentioning AI (up from 36 10 years ago, and 700 two years ago). Apparently this is what happens with technology trends and Internet of Things was mentioned in most a few years ago. An interesting point to note is the increase in blank check companies mentioning AI (note: these are "known as a special purpose acquisition company, is an early or development stage entity created specifically for the purpose of acquiring or merging with one or more existing businesses. It doesn't have a business model of its own beyond these financial transactions." according to Smart Asset) (Read more)

Nvidia Stock Has Risen More than 200% This Year

Speaking of quarterly reports, Nvidia have out performed their predictions. Nvidia create the hardware (specifically the chips) which enables a lot of the current AI stuff to be done. Therefore, it's a bit unsurprising that they are having a record year (Read more)


AI in Real Life

AI Allows Paralysed Woman to Speak

Advances in brain-computer interface (BCI) have allowed a paralysed woman to communicate using a digital avatar (Read more)

If you are interested, here is the academic paper on Nature.


Trending - AI and Schools

A lot of news articles, thought pieces and regulation have been announced in the AI and education space this week, particularly on school education. First a bit on the background of AI and education policy announcements in recent months:

  • UK Department of Education have just concluded a briefing on 23rd August about better understanding the role of AI in education. The results are expected in Autumn (Read)

  • US Office for Educational Technology have published a report, Artificial Intelligence and the Future of Teaching and Learning (Read)

  • UNESCO have published Generative AI and the future of education (Read)

  • European Digital Education Hub released a briefing on Use Scenarios & Practical Examples of AI Use in Education (Read)

These are just some examples which many countries and regions publishing their own findings and guidelines.

In the News This Week - Thought Pieces

It seems this week everyone has an opinion on the use of AI in schools.

The New York Times published two quite balanced pieces - How Do Teachers and Students Feel About AI and How Schools Can Survive (and Maybe Even Thrive) With A.I. This Fall. A similarly balanced overview was written (by the humans) of FE news.

The Times have gone with a slightly different approach Letting AI teach ‘is like using Casualty actors in A&E’ .

The focus in Business Standard and Wired have been around students cheating using these tools (although Wired's does end in a discussion around embracing them as tools, so not all doom and gloom).

The director of AQA (one of the exam boards in the UK) has given a written response to the Government's call for evidence which is a very useful read on the potential benefits (cutting teacher workload) and potential issues (AI requiring a greater need for exams) - read more.

Newly Published Research

A newly published paper shows promising results by putting MOOC (Massive Open Online Courses, eg Coursera) against an Intelligent Tutoring System (Read paper).

Further Reading

  • Dr Philippa Hardman is a world leading expert in AI & Education - she has released a news letter with lots of additional reading - read

  • We have already shared this on LinkedIn this week - but the episode of Possible where Reid Hoffman and Aria Finger speak to Ben Nelson founder of the Minerva University about the future of education


Until next week - if anyone sees anything they think we should include do not hesitate to send it across. Enjoy the bank holiday weekend if you're in the UK.

A representation of AI created by Deep Mind

12 views0 comments

Recent Posts

See All
bottom of page