Gov vs AI or Gov with AI?

Hey there, AI Enthusiast! 🤖

Remember me, reader? Back again, Max and my 3-minute newsletter. Before we delve into this week's news, I want to express my sincere gratitude for your continued presence: every one of you is extremely important to me. Thank you for reading, subscribing, and supporting me with your questions and reactions I receive ❤️. 

Aren't you just the absolute best? 🥹 Can’t wait to dive with you into the latest updates. Here we go!

1. Up-to-date government 📵

In anticipation of the forthcoming general elections, the UK government appears to have another concern on its plate. The 2024 Edelman Trust Barometer, an annual survey gauging public opinion on various institutions, indicates that most people in the UK 🇬🇧 lack trust in the government's capability to regulate artificial intelligence. According to the survey, two-thirds of respondents in the UK believe that the government lacks a sufficient understanding of emerging technologies to regulate them effectively.

The report reveals that 66% of individuals in the UK feel that "government regulators lack adequate understanding of emerging technologies," the highest percentage in the 28 countries surveyed (should we take pride in such a record? 😳). The lack of trust is particularly evident in the realm of AI, with 43% of respondents expressing hesitation to embrace AI-related innovations due to concerns about their management by both the government and corporations. In contrast, only 26% would be willing to embrace such innovations.

Critical concerns cited by respondents include issues like disinformation and job security associated with AI. Despite the UK government's efforts to position itself as a leader in AI regulation, hosting the world's first AI Safety Summit and signing the Bletchley Declaration with 28 nations, critics argue that these agreements do not adequately address existing AI risks. They emphasise the lack of specific legislation governing the technology in the UK.

Will it change? Let's purchase some popcorn and observe 🍿 Make or break, as they say.

2. May I talk to Mr.President? 🕵🏻‍♂️

Hey, I’m not a robot; thanks to the internet for checking in on me. But wait! There’s a new deepfake technology in town. Recently, a deepfaked President Joe Biden has been robocalling voters to discourage them from voting in the state of New Hampshire. Yikes 😬 

Researchers have also warned that fake audio files pretending to be politicians are likely to spread following instances of their being used in the UK, India, Nigeria, Sudan, Ethiopia, and Slovakia. They are becoming popular with the advent of cheap AI tools such as ElevenLabs, Resemble AI, Respeecher and Replica Studios. Microsoft has a new AI model, VALL-E, developed last year, which can clone a voice from three seconds of recordings(!). “There’s much less awareness about how audio material can be manipulated, so that, to me, really primes us to be vulnerable.” Henry Adjer is an expert on AI and deepfakes and an adviser to Adobe, Meta and EY.

Political experts are now raising the issue over viral deepfake audio clips as well as the use of deepfakes for robocalling or campaigns “You can very inexpensively build a strong, wide campaign of misinformation by phone-targeting people,” said AJ Nash, vice president at cyber security group ZeroFox. Experts warn that AI-created audio is often harder to detect than video 🥲. You can cover all the flaws of the recording with background noise and muffled music.

Online platforms are rushing to contain the issue, with Meta being criticised for not applying the same rules to manipulate video as it does to audio. Tiktok has also been investigating labelling and detection. 

Ensure the caller on the other end is who they say they are. Be safe out there, folks! 🙏🏻

3. Gemini Horoscope ♊️

Google is rolling out a conversational experience powered by its Gemini AI model in Google Ads, allowing advertisers to generate relevant ad content, including creative and keywords, from a website URL. It was announced as the first model to outperform human experts on MMLU (Massive Multitask Language Understanding), one of the most popular methods to test AI models' knowledge and problem-solving abilities. Announced last May, the conversational experience is now available in beta to English-language advertisers in the U.S. and U.K., with a global rollout in the coming weeks.

The tool has proven particularly beneficial for small business advertisers, with a 42% higher likelihood of publishing search campaigns with "Good" or "Excellent" Ad Strength. Advertisers who improve ads from "Poor" to "Excellent" experience an average 12% increase in conversions. Responding to concerns about image creation, Google plans to introduce a feature suggesting images using generative AI and images from landing pages, with watermarks for transparency 🌌

Generative AI is poised to impact Google and revolutionise the advertising landscape profoundly. Shockingly, an estimated 84% of Google Search queries are anticipated to be supercharged by generative AI, potentially unleashing a tidal wave that could result in over $40 billion in annual ad revenue, as revealed by data from the marketing platform BrightEdge shared with Marketing Dive. The industries set to experience the most transformative effects are healthcare, e-commerce, and B2B technology. Brace yourselves for a massive shift in the advertising realm! 🏎️

And with that, we conclude today's update. I trust it piqued your interest as much as it did mine! 🚀

Catch you in a week (or less) —stay curious! 🦾

xx

Max