top of page
  • Writer's picturelauralikespi

AI News - Fri 15th Sept 2023

Yet another busy week in the AI world - not a lot of real life applications being announced, but a large number of investments in a range of companies (from $5.1 million to $500 million investments). There have also been several new features in well known products , reports and surveys realised. The biggest news of the week is the US Senate sessions on AI.


 


Good to Know


US Senate - Hosts Three AI Sessions in A Week


It has been a busy week in the US Senate in terms of AI with three sessions taking place - two public and one private.


Private AI Safety Forum


A group of companies were invited to discussed AI with the US government this week. The session has been called historic and all reports suggest it went well. The aims were to discuss how to safeguard against AI, particularly with regards to misinformation in the 2024 elections.


All attendees (taken from here)

  • Sam Altman, CEO of OpenAI (developer of ChatGPT)

  • Rumman Chowdhury, CEO of Humane Intelligence

  • Jack Clark, Co-founder of Anthropic

  • Clément Delangue, Co-founder and CEO of Hugging Face

  • Eric Fanning, President and CEO of Aerospace Industries Association

  • Bill Gates, Former CEO of Microsoft and Co-chair of Bill & Melinda Gates Foundation

  • Tristan Harris, Co-founder and ED of Center for Humane Technology

  • Jensen Huang, President and CEO of NVIDIA

  • Alex Karp, Co-founder and CEO of Palantir

  • Arvind Krishna, CEO of IBM

  • Janet Murguía, President of UnidosUS

  • Elon Musk, CEO of Tesla and Chairman of X (formerly Twitter)

  • Satya Nadella, CEO of Microsoft

  • Sundar Pichai, CEO of Google

  • Deborah Raji, Researcher from University of California, Berkeley

  • Charles Rivkin, Chairman and CEO of Motion Picture Association

  • Eric Schmidt, Former CEO of Google and Chair of Special Competitive Studies Project

  • Elizabeth Shuler, President of AFL-CIO

  • Meredith Stiehm, President of Writers Guild of America West

  • Randi Weingarten, President of American Federation of Teachers

  • Maya Wiley, President and CEO of Leadership Conference on Civil and Human Rights

  • Mark Zuckerberg, CEO of Meta (parent company of Facebook, Instagram and WhatsApp)

There was apparently (according to Elon Musk) "consensus" and "all we needed was a referee". Although no explicit plans or time scales for what the group agreed has been released.


Read some more news coverage - The Guardian and BBC.



Public Subcommittee Sessions -

Oversight of A.I.: Legislating on Artificial Intelligence - this was a bipartisan session lead by Chairman Sen. Richard Blumenthal (Democrat representative for Connecticut) and Ranking Member Josh Hawley (Republic representative for Missouri). A law academic from Boston Unviersity, Chieft Scientist of NVIDIA and Vice Chair of Microsoft were the witnesses. The topics discussed included licensing agreements, kids privacy and national security (Watch session - transcript)


Another session also took place, Governing AI Through Acquisition And Procurement (Watch session)



Bristol, Our Home City, to Host Supercomputer to Power AI Innovation

The UK's most super supercomputer, Isambard-AI, will be built in Bristol later this year. This computer which will also be one of the most powerful in Europe comes with a £900 million investment to drive AI innovation in the region (Read more)



California Senator Suggests AI Rules Should be State Level

Despite the goings on at a country-wide level in the senate (mentioned above) because “California has a history of acting when the federal government is moving either too slowly or not acting.” This move makes sense with Silicon Valley, California being the home of a lot of AI innovation. The bill focuses on transparency and security of Large Language Models, however it is noted (as with every other attempt) this bill is little light on details (Read more)



Google is Hoping $20 Million Will Solve Some AI Issues

Google has launched a new initiative, Digital Futures Project, which aims to fund academics and think tanks to explore important themes within AI. The themes they give as examples are:

  • How will AI impact global security, and how can it be used to enhance the security of institutions and enterprises?

  • How will AI impact labor and the economy, what steps can we take today to transition the workforce for AI-enabled jobs of the future, and how can governments use AI to boost productivity and economic growth?

  • What kinds of governance structures and cross-industry efforts can best promote responsible AI innovation?



Amazon Have Added New AI Declaration for Self-Publishing

When an author adds or edits a book on the Amazon self-publishing platform, they must declare if any of the content (including artwork) is AI-generated. They do go on to distinguish between AI-generated and AI-assisted. The latter, which includes when the author "used AI-based tools to edit, refine, error-check, or otherwise improve that content (whether text or images)" or for brainstorming, does not need to be declared.



LinkedIn Published Findings From Future of Work Report: AI at Work Report

LinkedIn have published the findings from speaking to 29k people from various countries about AI at work. We love this quote:


Nine out of 10 professionals are excited to embrace AI, but half admit they need to know more.


Some other interesting findings:

  • AI skills - on LinkedIn there are 38k skills, 121 of these are AI skills including Alexa and natural language.

  • Marketing and sales use AI a lot - 69% of sales people expect to use AI in their job in the next 6 months

  • Jobs being replaced - 96% of software engineering tasks can be completed by GAI, 45% of teachers and 6% of nurses

The report is worth a read, including the full 121 AI Skills (Read more)



KPMG Report - Privacy in the New World of AI

This KPGM reports looks at how regulations are evolving as GenAI evolves, and it includes a good overview of how nations are trying to regulate AI. They set out nine principles for AI privacy:

  1. Lawfulness and fairness

  2. Transparency and explainability

  3. Governance and accountability

  4. Data-minimization

  5. Purpose-limitation

  6. Accuracy

  7. Storage-limitation

  8. Security

  9. Respect for end user privacy



EY is Investing $1.4 Billion in AI

EY (Ernst and Young) want to help their staff use AI effectively, so are using $1.4 billion to roll out an AI platform (EY.ai), their own Large Language Model (EYQ) and an upskilling programme for their staff (Read more)



Alibaba new focus on AI

Alibaba, the global giant, has announced a new CEO and a big focus on becoming "AI-driven". They have also said they are aiming to hire more young people (apparently this means people born in 1995 and after) (Read more)



Snowflake Says AI Usage is Up 15% in Marketing Teams

Snowflake, a cloud-computing data company, have released a report Modern Marketing Data Stack 2023: How Data-Forward Marketers Are Redefining Strategies to Unify, Analyze, and Activate Data to Boost Revenue. As well as stating marketing teams use of AI is up 15% year on year, they identify the three data categories with the highest growth - Measurement & Attribution, Customer Data Activation, and Analytics & Data Capture. Their report is overall highlighting how much data and AI are becoming important to marketing teams (Read more)



Conference Season Kicks Off

Generative AI in Education Conference - A conference, GenEdu, brought together academics and industry experts to discuss Generative AIs uses in education in the Northern Irish city of Derry (Read more)


Forbes Global CEO Conference - At the annual conference, this year in Singapore, AI was the most talked about topic. The overall tone seemed optimistic but cautious of the potential harms (Read more)

 

AI in Real Life


AI Can Detect Health Conditions by Scanning Eyes

A very interesting application of approaches used in Large Language Models, such as those behind ChatGPT, have been used to detect a range of diseases including heart disease and Parkinson's, by scanning the retina of eyes. "RETFound uses a multitude of retinal photos to learn how to predict what missing portions of images should look like". (Read more)


Coke Brings Out a New Flavour - Designed by AI

Channeling their inner Busted fan girl, Coke has launched a new flavour - Y3000. Coke describe this as a flavour from the future "co-created with artificial intelligence" (Read more)



Salesforce Introduces Einstein 1 Platform

A pilot has been launched by Customer Resource Management giant, Salesforce. The product which they are calling Einstein 1. The aim is to use Generative AI assistants seamlessly within their already existing workflows (Read more)



Adobe Has Launched AI Tools for Photoshop

Adobe has announced its new generative AI tools - Generative Fill (to add things to images) and Generative Expand (to edit images) powered by Adobe Firefly (Read more)



AI is Being Used in Timber Construction

An Italian company, CRA-Carlo Ratti Associati, are pioneering a new approach to timber construction. They use AI and LiDAR technology to minimise waste and create innovative designs (Read more)




AI Girlfriends Are Popular - Who Would Have Thought?

A chatbot company which allows you to speak to your AI girlfriend, Dream GF, has reported over 500k users, 100k paying users and $120k revenue per month. The website is exactly what you would expect - beautiful women with faces that are potentially obviously AI-generated but no one will be looking at their faces to even notice. For more money, your girlfriend will send you some pics (sigh). The company told Sifted they plan to speak to OnlyFans stars about using their "appearance and personality". Our usual line about "there is literally a Black Mirror episode about this" definitely applies here (Read more)


 

Companies to Watch

There have been so many investment announcements this week, we have ordered them from largest amount to smallest amount.


Databricks, an All-Encompassing Data Lake, Have Raised a Series I Round of Over $500 Million

Databricks, a company we love, who have been pivoted in many companies' data journeys are now valued at $43 billion. This follows their latest investment raise which was $500 million for a Series I round (Read more)



Generate:Biomedicines, AI-Drive Drug Discovery, $273 million in a Series C

Generate:Biomedicines who use AI to development treatment for various diseases, including COVID, have raised a Series C round of $273 million. Their backers include NVIDIA. They will use this investment to "increase significantly the precision, speed, and probability of success of novel therapeutics – positioning us to tackle even more complex targets and unmet patient needs" (Read more)



Helsing "Defense AI" Raises €209 Million in Series B (Wow)

Helsing (do check out their very slick website) have closed their enormous Series B round headed up by General Catalyst. Even the wording of their mission "to bolster defence AI capabilities for democracies" makes them genuinely seem like a defence company with a difference, and it has obviously paid off as this makes them the largest European defensive unicorn (according to TechCrunch). They have been chosen as a partner for the German government on an "AI-enabled electronic warfare capabilities", and the Spotify founder was lead on their Series A round (Read more)



Enfabrica, an AI Chip Manufactor, Have Raised $125Million in Series B

Enfabrica have upped their valuation (by 5x) in their latest fund raising round. They will be using the money for their R&D on networking chips aiming to make the computer speed faster and memory usage smaller. This round included backing from NVIDIA (who also manufactor AI chips) (Read more)



Pixis, an AI Marketing Platform, Raise Again - Series C1 of $85 Million

Pixis is an all encompassing AI solution for marketing - with three products Targeting, Creative and Performance. They have raised for a second time this year, following a $100 million Series C in January. Despite have over $50 million in annual recurring revenue, Pixis are raising this round to focus on R&D (Read more)


Druid, Conversational and Gen AI for Microsoft Businesses, Closes a Series B Round of $30 Million

Druid aims to "enhance enterprise productivity through conversational business applications and Generative AI". They do so specifically on Microsoft Azure platforms (Read more)



LastMile AI, an AI-Platform for Developers, Raise a $10 Million Seed

A group of developers from Facebook / Meta AI who built well known internal developer tools for the company are building an AI-powered platform for developers, LastMile AI. They have closed a seed round of $10 million led by Gradient Ventures (Read more)



Databutton, AI-Power for Everyone, Raise a Seed Round of $5.1 Million

Databutton, an Oslo based company, who want to help everyone build chatbots and other AI-powered assistants with the click of a button have raised a seed round (Read more)


 

Interesting Reads

  • Excellent visual description of how transformers (which we can thank for how well Large Language Models work) from the Financial Times

  • Wired interview with Google CEO, Sundar Pichai, on the company's 25 anniversary

  • Wired article on the prisoners being used to label data for AI (and the country will surprise you)

  • Another Financial Times article about AI impending impacts on politics and society (while promoting three books from experts on AI)

  • IEEE article on whether we are ready for the chatbots currently joining our workforce

 

4 views6 comments

Comments


bottom of page