The SEC Newgate AI Weekly: Keeping you up-to-date on the latest developments in AI this week

It’s been a week since we heard about the open letter pleading for a six month pause in AI development, citing concerns towards society and humanity – a request that has 41% of US adults strongly supporting in a recent YouGov America poll.

While this plea to decelerate the pace of AI development could suggest that cautionary tales (or Hollywood blockbusters) like ‘Bladerunner’ or Mary Shelley’s ‘Frankenstein’ are playing on the subconscious minds of tech leaders, the more cynical mind may see it differently, rather as a plea by the likes of Elon Musk to deny other AI tools the time to catch up to the likes of GPT5, scheduled to launch in six months… time will tell which of these is the case.

Since then, it seems conversations around AI have been propelled not only to the mainstream arena but as reported by Reuters, are also becoming a ‘hot topic for policy makers’ as Biden draws in on AI and the role tech companies have in ensuring product safety before AI tools are unleashed to the world.

Two papers were released last week by AI researchers Sam Bowman and Dan Hendrycks,: ‘Natural Selection favours AI over humans’ and ‘Eight things you should know about large language models.’ Significantly, one of the key findings from the latter report found that already there are ‘no reliable techniques for steering the behaviour of Large Language Models (LLMs).’ Perhaps a reason for the six month pause, if we’re already surpassing the ‘building’ stage and entering the ‘steering’ phase?

In other AI news, Midjourney, the tool that allows users to create fake images, such as the Pope in a Balenciaga puffer jacket and Trump running from police, has now removed access to its free trial after the explosion of deep fake imagery, reports Decrypt. Interestingly, Midjourney has also banned users from developing images of Chinese President, Xi Jinping – a move that has critics questioning how Chinese censorship impacts overall ‘free speech and expression.’

Additionally, in response to a Washington Post report that criticised Snapchat’s attempt at creating a ‘safe AI chat box,’ the platform is planning to launch new safeguards to combat inappropriate conversations with the chatbot just three weeks later, as reported by Tech Crunch. While these measures by the two example companies are well intended, the pace of AI development suggests that any sort of regulation will remain almost a pipedream.

Yesterday, AI-powered-voice was making waves on Twitter, with multiple videos providing examples of AI voice generation and users hypothesising the potential repercussions toward the entertainment industry.

What else?

1. Meta releases AI model that can identify items within images reports Reuters.
2. Forbes has reported the ChatGPT of Finance is here with Bloomberg combining Fintech and AI.  Known as BloombergGPT.
3. While news has been saturated by other tech giants racing to generate AI tools, there was a notable silence from the likes of Meta. This was until yesterday when they announced a plan to use generative AI to develop ads by end of year, reports Tech Crunch.
4. ‘Google reveals its newest A.I. supercomputer’ reports CNBC
5. Fashion giants Levi’s and Calvin Klein have utilised AI to create ‘diverse AI models’ reports the Guardian.