top of page



Preparing for AI

Your weekly CogX newsletter on AI Safety and Ethics
The week's developments on AI, safety and ethics, explained | 26.04.24

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

This week the UK cracked down on non-consensual deepfake creation, banning a sex offender from using any AI tools for 5 years (with much stricter legislation in the works) — the move has prompted two prolific explicit AI websites to block UK access. Will other countries follow suit?


Meanwhile, Meta’s AI was let loose on Facebook (it started lying about having kids), Sunak wants to replace GPs with algorithms to mitigate ‘sick note culture’ and OpenAI are trying to control India’s tech regulation through strategic hiring.


We cover this, plus the UAE approach to regulation, what early philosophy can teach us about AI and Microsoft’s deepfake dilemma, in our Preparing for AI weekly digest.

- Charlie and the Research and Intelligence Team

P.S Limited availability for our CogX Transatlantic Accelerator at $795 (RRP $23,621). If you are or know a UK startup, they can apply now for over $20k worth of support to attend, exhibit, and network at the CogX Festival in LA on 7th May here.


Share your expertise! Want to be a guest contributor in our next issue? drop us a line at:

Ethics and Governance


🏛️ Landmark UK case bans a sex offender from using AI tools for five years after generating over 1,000 indecent images. This decision follows recent legislation that, whilst not yet passed, will threaten ‘despicable’ offenders with fines and potential jail time.


🚫 Two largest deepfake pornography websites blocked in the UK, following government crackdown on non-consensual deepfakes. Whilst websites can still be accessed by VPN, the UK’s move towards criminalising the creation of explicit AI material remains significant.


🇮🇳 OpenAI hired a head of government relations in India to shape regulation early amid elections. The move shows OpenAI's proactivity to foster beneficial AI development in India, as the country emerges a critical market for global tech advancements. 


🇦🇪 What can Britain learn from the UAE’s ambitious AI policy? Termed a 'regulatory sandbox' by Altman for its lenient approach, the UAE allows AI training on government data. Britain could adopt a similar strategy, particularly with NHS data, to boost AI development.

📝 NHS algorithm will write sick notes instead of doctors in a UK initiative to reduce GP workload. PM Sunak emphasised the need to change prevailing "sick note culture" to focus more on the tasks individuals can perform. Details on the tech’s specifics are not yet confirmed.


AI Dilemmas

🤖 Microsoft’s VASA-1 can deepfake anyone from just one image and audio clip, with insane realism. The cutting-edge AI can match facial expressions, head movements in lip-synching perfectly with the audio, raising concerns over misinformation and misuse.

🤰 Meta's AI claimed it had a disabled child in a gifted and talented program in a private parenting Facebook group. The new AI automatically responds to posts unanswered for an hour. Whilst hallucinations are commonplace, should they be tolerated on social media?

🚀Enjoying our content? We’re also on LinkedIn — Follow us to stay in the loop with the week's most impactful AI news and research!


Insights & Research

🛡️ AI is the future of national security decision-making, finds Alan Turing Institute in a new report, due to its extensive data processing and beyond-human pattern recognition. The institute also stresses the need for specialised training in national security to mitigate risks. 


💰 The AI race presents a dual reality where tech giants lead in research and investment, but also face rising costs. While they haven't yet monopolised AI, leaving space for emerging players and economies, concerns linger about potential techno-feudalism or datacolonialism.


📚 What early Islamic philosopher Ibn Sina can teach us about AI, specifically whether it deserves digital personhood. Despite living years before the invention of the printing press, let alone AI, his research into what makes a human, human, is particularly relevant today. 

📈 How to mitigate harms from persuasive gen-AI: As the tech becomes more persuasive and prevalent, it becomes more concerning. This paper distinguishes between rationally persuasive and manipulative AI, outlines potential harms, and proposes how to mitigate it.

In case you missed it

'What does it mean to make something totally new, fundamentally different to any invention that we’ve known before?’ — Mustafa Suleyman grapples with what AI was, is, and will be in the future:

✍️ 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

🚀 CogX Transatlantic Accelerator Early Bird offer closes on 15th April: If you are or know a UK startup they can apply for over $20k worth of support to attend, exhibit and network at CogX Festival in LA on 7th May here

bottom of page