top of page

Preparing for AI

Your weekly CogX newsletter on AI Safety and Ethics
The week's developments on AI and regulation, explained | 05.01.24

If you’re enjoying this briefing, sign up here — and share it with a friend. Reading time: ~5 minutes


As the US gears up for elections, the potential of gen-AI to spread misinformation may put democracy under threat. Experts are calling for enhanced support and stringent security measures from federal authorities. Across the pond, a UK anti-extremism think tank urged the government to develop anti-AI terrorism laws, as the pace of development will outgrow current legislative measures by 2025.

 

Meanwhile, in healthcare, the rising use of ChatGPT for self-diagnosis signals the importance of ensuring AI tools support, rather than undermine healthcare. Plus, could AI prompts be patented?

 

We also cover the experts’ take on the job market, EU AI Act, plus the latest research on AI creativity, and LLM psychologists.

 

- Charlie and the Research and Intelligence Team


Ethics and Governance



🗳️ AI is a direct threat to democracy — here’s how to handle the US elections. With elections approaching, the rise of gen-AI poses a serious threat through misinformation. Addressing this requires robust support for state and local election officials from federal bodies, alongside stringent security protocols and cross-sector collaboration. 

 

🚨 We need anti-AI terrorism laws urgently, warns UK think tank. AI is predicted to aid violent actors as soon as 2025. With concerns about misuse on the rise, and current policy failing to address impending threats, the call for legislative updates and increased safety measures grows stronger. 

 

📰 New York Times sues OpenAI and Microsoft for copyright infringement. The lawsuit seeks to halt the use of The Times's copyrighted content in AI model training, and addresses broader concerns about AI's impact on journalism and content ownership. 


🏛️ 2023 US AI Policy Recap - what’s next for 2024. From privacy to authenticity, 2023 was a huge year for AI policy. So, what can we expect from 2024? Tech Policy Press predicts standardised AI content regulation and comprehensive AI risk assessment, plus robust government funding and implementation.


 

AI Dilemmas


🔍 Can AI prompts be patented, and if so, why? AI prompts may offer broader intellectual property protection than to the outputs they produce. For a prompt to qualify for a patent, it needs to be both novel and useful. The main debate is about how specific these prompts must be, and how they might impact innovation. 


🩺 The increasing role of ChatGPT in self-diagnosis. Chatbots offer unprecedented access to medical information, helping people better understand their symptoms. However, as patients increasingly turn to AI for health insights, the medical community must ensure this tool complements, rather than compromises, the quality of healthcare.


 


Insights & Research



🎨 Can AI be as creative as humans? This Turing test-inspired framework, coined "Relative Creativity" may help us find out. The framework compares AI creative output to hypothetical human output, and assesses if the work is indistinguishable. It hails a practical, comparative approach, rather than trying to define creativity in absolute terms. 

 

🇪🇺 I advised the EU AI act - here’s what I learned. David Haber discusses how the Act – originally designed for a narrower AI scope – was challenged with the surge of generative AI, and adapted through a gen-AI addendum. He emphasizes that the act is still evolving, with key technical and industry controls pending in 2024. 

 

💼 Does AI signal the end of work? Ro Khanna doesn’t think so. He believes we can — and must — balance AI development with the security of the working class. Khanna advocates for workers' involvement in decision-making and profit-sharing, proposing a balanced approach to innovation that prioritizes workers' well-being.

 

🧠 Can we use LLMs in psychology? This research paper evaluates the myriad benefits, ranging from enhancing literature reviews to assisting in experimental design and data analysis. However, it also highlights the technical and ethical challenges of using LLMs in psychology, emphasizing the importance of responsible use and design.


In case you missed it


Confused about the AI that supposedly got Sam Altman fired? Check out the full explanation here:



✍️ Enjoying this newsletter?  Subscribe to our free weekly briefings on Preparing for AI, Cinema & AI, The Race to Net Zero, AI & DeepTech and The Future of Work.

bottom of page