AI Weekly Digest - October 23 2023
Welcome to our AI update, we're delighted to announce that following the enthusiastic response to our LinkedIn newsletter we have launched an enhanced email edition of the AI Weekly Digest.
You can subscribe for free and benefit from:
IDC: Global Gen AI spend to climb 73% a year to 2027, to hit $143bn
IDC said it expects global spending on generative AI solutions to grow at a compounded average of 73.3% in the five years to 2027, when the segment will be worth $143bn. Spokeswoman Ritu Jyoti, said: "Generative AI is more than a fleeting trend or mere hype. It is a transformative technology with far-reaching implications and business impact. With ethical and responsible implementation, GenAI is poised to reshape industries, changing the way we work, play, and interact with the world."
CEO: ITV trialling using AI in post production workflows
ITV CEO Carolyn McCall said the commercial broadcaster is trialling using AI in post production workflows. She told the Hollywood Reporter that she sees generative artificial intelligence offers a “significant opportunity”, adding: “Although it is easy to see the threat of generative AI, and some of that can be overwhelming, we are doing everything we can to explore and experiment so we are clear where to focus our efforts - on the opportunities but also on mitigating the threat.”?
Clegg: Meta to stop 'inappropriate’ replies from chatbots
Meta?president of global affairs Sir Nick?Clegg said the group has put in "thousands of hours of red teaming - stress testing" to make its new generative AI bots "as safe as we possibly can", but conceded "we can't guarantee that they won't spew something inappropriate or something unwelcome at any point”. He told an interviewer: "The more open we can be about how these new technologies operate, the more open we can be about what the risks and the benefits are, what the controls are for users and also how we need to work with others."
领英推荐
OpenAI’s GPT-4 more reliable but vulnerable to 'jailbreaking'
Microsoft-backed research showed that while OpenAI’s GPT-4 large language model is more reliable than previous versions, it is also more vulnerable to 'jailbreaking' when told to ignore security measures and leak personal information. The research team said: “This trustworthiness assessment is only a starting point, and we hope to work together with others to build on its findings and create powerful and more trustworthy models going forward.”
Baidu unveils Ernie Bot 4.0
China's Baidu has unveiled the latest version of its generative artificial intelligence model, Ernie 4.0. It said the AI bot's capabilities were on a par with?ChatGPT-maker OpenAI's GPT-4 model. The tech giant's CEO Robin Li unveiled Ernie 4.0 at an event in Beijing, highlighting the model's memory capabilities and demonstrating it writing a martial arts novel in real-time. He also showcased Ernie 4.0 creating advertising posters and videos.
Report calls for 'meaningful' UK legislation to become global AI leader
The UK government must create “meaningful” legislation for this country to become a global leader in AI, a Cambridge University report has found. It stated that without “solid legal and ethical AI regulation” the public and industry may struggle to trust emerging home-based artificial intelligence platforms and organisations’ use of the technology. The report comes ahead of the first international AI safety conference, which will be hosted by the government in November, at Bletchley Park. Professor Gina Neff, executive director of the Minderoo Centre for Technology and Democracy, said: “A new bill that fosters confidence in AI by legislating for data protection, intellectual property and product safety is vital groundwork for using this technology to increase UK productivity."
Meta chief AI scientist: Regulating R&D 'incredibly counterproductive'
Meta’s chief AI scientist, Yann LeCun, warned "regulating research and development in AI is incredibly counterproductive", telling the FT ahead of next month’s Bletchley Park conference that current market leaders "want regulatory capture under the guise of AI safety".
Report: German?chancellor set to miss key UK summit
The BBC said it understands German chancellor Olaf Scholz could turn down an invitation to the AI safety summit the UK is hosting at Bletchley Park next month, prompting concerns that other world leaders could follow suit.