A week in Generative AI: Emotions, self-regulation and generative expand
A week in Generative AI: Emotions, self-regulation and generative expand
News for the week ending 30th July 2023
This week we saw the White House secure voluntary commitments from the major AI labs to self regulate, a new report into the devastating water usage of data centres, and the release of Photoshop generative expand which works on both images and videos. What a week!
Self-regulation of AI
Last week, the major AI labs met at the White House and agreed to establish the ‘Frontier Model Forum’ - a new industry body to promote the safe and responsible development of frontier AI systems.
Companies that signed up are Amazon, Anthropic1, Google2, Meta, Microsoft3, and OpenAI4 and they agreed to eight measures covering areas such as Safety, Security and Trust:
Using watermarking on audio and visual content to help identify content generated by AI.
Allowing independent experts to try to push models into bad behavior – a process known as “red-teaming”.
Sharing trust and safety information with the government and other companies.
Investing in cybersecurity measures.
Encouraging third parties to uncover security vulnerabilities.
Reporting societal risks such as inappropriate uses and bias.
Prioritizing research on AI’s societal risks.
Using the most cutting-edge AI systems, known as frontier models, to solve society’s greatest problems.
Thirsty Data Centres
A new report from Bloomberg highlights how battles over water are erupting between AI companies seeking more computing power and local communities where their facilities are located. The report centres on Spain where droughts have sent water levels in dams below historical averages and leading to restrictions being put in place to guarantee water supplies during the day.
“People are not aware of the amount of water that goes into watching a kitten meme.”
The report goes into details about how the hardware used to train generative AI models requires more powerful specialised computer processors that emit more heat than general-purpose processors and so require much more water to keep cool.
Emotional AI
In a talk at King's College in London, Geoffrey Hinton expressed another thesis that is likely to stir emotions in the AI industry.
When asked is AI systems could one day exhibit emotions, Hinton said":
"I think they could well have feelings. They won't have pain the way you do unless we wanted, but things like frustration and anger, I don't see why they shouldn't have those."
This is the first time that Hinton has publicly expressed his belief that machines could develop emotions.
Photoshop Generative Expand
Adobe released a new Generative Expand workflow in Photoshop that generatively fills both images and videos.
“The future is already here, it’s just not evenly distributed.“
William Gibson
You can also subscribe to some of the other Blueprint Newsletters to stay up-to-date with Generative AI:
The Blueprint - A series of articles that share actionable and agile strategies for implementing Generative AI.
The Classroom - An education series on Generative AI.
The Soapbox - Sharing great resources and shouting about important topics in Generative AI.
The School Yard - Where I kick around and discuss interesting concepts in Generative AI.
Resources - A page of resources that I’ll keep up-to-date as I see useful and interesting things related to generative AI.
This article was researched and written with help from ChatGPT, but was lovingly reviewed, edited and fine-tuned by a human.
https://www.anthropic.com/index/frontier-model-security
https://blog.google/outreach-initiatives/public-policy/google-microsoft-openai-anthropic-frontier-model-forum/
https://blogs.microsoft.com/on-the-issues/2023/07/26/anthropic-google-microsoft-openai-launch-frontier-model-forum/
https://openai.com/blog/frontier-model-forum