Below is a list of article links with some comments to summarize their content
Artists file lawsuit against AI content generators for misusing copyright
- Stability AI, mid-journey, and Deviantart are being sued for copyright infringement, by collecting art data without permission.
- Getty Images takes legal action against Stability AI
- A class action lawsuit was also filed against Github and OpenAI for using copyrighted source code without permission.
US Senate press release going over plans for bipartisan AI briefing
- Will discuss how AI is currently used and what it is capable of
- Will discuss where AI will develop in the next decade.
- Will look into how DoD uses AI and how enemies of the US use AI.
CEO of OpenAI, the creator of Chat Gbt, approaches congress with concerns of AI
- Lawmakers are impressed with how upfront the CEO (Sam Altman) of OpenAI is (Chat Gbt)
- Altman states that jobs will become replaced by AI, but will be replaced by better jobs. He seeks government support to mitigate the risks of job loss
- People are wary about how willing Congress is to listen to how a CEO would regulate his own tech.
- Altman was direct with the various risks present with the use of AI technology (I.E., AI-piloted drone strikes, job automation, mass manipulation of information)
New laws the EU plans to pass for AI regulation
Will ban the following uses of AI
- Realtime and post-remote biometric ID in public spaces
- Exception for post systems for use in serious crimes with judicial authorization
- Predictive policing systems
- Emotion recognition in law enforcement
- Sourcing data from social media/CCTV to create facial databases
- Allows researchers to be innovative by lifting restrictions on research and open-source projects
- Apply transparency - Disclosing content was made by AI, preventing illegal content generation, and summarizing copyrighted data used for training.
- Overwhelming support with 84 in favor, 7 against.
- Votes on proposed changes will happen on the June 12-15 session
Ethical AI Database methodology
- A database of startup companies that account for AI ethics
- Creates reports based off of changes in the AI ethics landscape
- 5 types of companies that are considered
- Data for AI
- Detect AI bias
- Ethical collection of AI data
- Maintain privacy
- ModelOps (https://en.wikipedia.org/wiki/ModelOps)
- Monitor the performance of AI
- AI quality assurance
- Blackbox explainability
- AI Audits (Risk/compliance) https://www.isaca.org/resources/news-and-trends/newsletters/atisaca/2020/volume-26/auditing-guidelines-for-artificial-intelligence
- Assess legal, ethical, and privacy risks associated with AI
- Assess organizational model risk
- Help advice for AI governance gaps
- Targeted AI solutions
- Try to solve specific problems caused by AI
- Develop a more ethical way of doing things (Advertizing, facial recognition)
- Also solve problems not caused by AI (Toxic content moderation)
- Open source
- Provide easy access to AI solutions
- Most are non-profit organizations
- Provide solutions to make AI more ethical
- Data for AI
Recent changes in US copyright law to account for the use of AI generated content
- The US copyright offices’ view of AI
- AI content can not be copyrighted, as it is not made by a human
- Humans can use AI content within copyrighted works, but the AI content itself will not be copyright protected
- For example, a book written by a human, using illustrations by an AI.
- The work on top of AI-generated content must be transformative in order to be copyrighted, where the AI is merely an assistant and not a replacement for creating content.
- The content generated by an AI within that content can still not be copyrighted
US white house published an over 70 page document called the “AI bill of rights”