The SEC Newgate AI Weekly

It’s been a hot topic in many party conference talks and conversations over the past week or two – for better and for worse. Many recent media stories also look at the emerging economic and environmental impacts around the energy required to power AI. But we start this week with some exciting news for creators and designers. 

Many tech publications have covered Adobe’s Firefly Image 2 Model this week. The imaging model has been updated with more knowledge and better technical capabilities, enabling it to generate hugely improved images compared to its predecessor. The comparison graphic in Adobe’s own article shows the difference in outputs. 

Kier Starmer fell victim to “UK politics’ first deepfake moment” on the first day of the Labour Party Conference, as reported by Politico. The fake audio emerged on Sunday morning, on an X account with less than 3,000 followers. It claims to capture the Labour leader abusing party staffers, but the audio is not real, and the incident did not happen, explains this Sky News clip — which also includes the audio. 

Talk continues about the government’s work in the lead-up to the UK’s AI Safety Summit at Bletchley Park on 1-2 November. Drafts of statements appear to refer to an “AI Safety Institute”, according to The Guardian. 

The BBC reported on a new study about AI industry energy consumption, written by a PhD candidate at the VU Amsterdam School of Business and Economics, explaining how the AI industry could consume as much energy as a country the size of the Netherlands by 2027. The Verge looks at Google’s use of AI specifically, quoting the study’s author in an interesting article. 

As well as environmental impact, this CNBC article looks at how costs might stall development of generative AI by small businesses in 2024, relative to the hype and speed of progress we’ve seen in 2023. 

FT Communities editor Lucy Warwick-Ching breaks down the current considerations for investing in AI in this piece. 

The Next Web published this interesting piece about start-ups on a mission to reduce generative AI hallucinations. Something we are talking to clients about more at SEC Newgate, hallucinations are both a pain for users looking for accurate information, whilst also posing a reputational risk for personal and corporate brands alike. 

While AI-enhanced profile photos have been a thing for a while now and continue to be the basis of many TikTok trends and filters, this BBC story explores how professionals are using AI-generated headshots for job applications. 

And finally, DIY AI article of the week goes to Surrey Live, which reported about a man who created an AI device that deters foxes and badgers from his garden — using machine learning after feeding it 800 images of the animals.