This website uses cookies to improve user experience and analyze website traffic. Privacy policy
Insight

Top AI headlines from this summer

The AI development shows no signs of slowing down. On the contrary, while you were hopefully enjoying some relaxation in a hammock, AI news has been unfolding rapidly. This summer, it became evident that generative AI is here to stay, the AI race among the tech giants continues, the use of AI in journalism is on the rise, and an increasing number of Swedish companies are embracing AI technology. Here, we list some of the most important AI news from the summer of 2023.

Image: Midjourney/Hanna Berglind

Generative AI Takes Center Stage

McKinsey QuantumBlack released its annual survey, “The State of AI”, which this year highlights the explosive growth of generative artificial intelligence. Even though many of the popular tools in generative AI were launched less than a year ago, a third of the respondents say their organizations use generative AI regularly in at least one business function. According to McKinsey, generative AI has the potential to account for 4.4 percent of the global economy's annual growth.

The growing interest in generative AI was also confirmed during Amazon's Q2 Conference when the company's CEO, Andy Jassy, announced that every team at Amazon is currently working with generative AI. He described the technology as a significant investment and a central focus for the company moving forward. Jassy emphasized that the technology will both help Amazon streamline costs and optimize its processes. In August, Amazon announced they will be launching AI Summaries of product reviews so consumers won't have to sift through all of the reviews. They have also previously mentioned that they're using AI to identify fake reviews.

Earlier this summer, Meta's research team, Audiocraft, launched the AI-based music generator, Musicgen, which can produce music based on a text description. The open-source generative model has been trained on 20,000 hours of licensed music, and users can now test the tool to generate a 12-second music clip.

Image: Midjourney/Hanna Berglind

Stability AI, the company behind the popular text-to-image generator Stable Diffusion in 2022, has now released a new version - Stable Diffusion XL. This model is claimed to be even better than its predecessor at creating photorealistic images and excels in areas that have been notoriously challenging for generative models, such as hands and accurate spatial relationships in images. Users can also upload images to SDXL, and the system can modify them according to user instructions.

Generative AI is also making significant strides in web development. The website company Wix launched an AI tool for creating web pages. Using generative AI, users can describe the type of webpage they want, and the system creates the website, which can include text, images, and specific sections for events, bookings, and more.

The tech giants' AI race is accelerating

News about the tech giants and AI seems never-ending. YouTube has announced that they've begun testing AI-generated automated dubbing for YouTube videos. Hundreds of content creators have been given access to this feature, but YouTube hasn't disclosed if or when a broader rollout will occur.

Google (Alphabet), the owner ofYouTube, is also advancing AI efforts in other areas. During the summer, for instance, they launched two new AI tools for advertising. Demand Gen aims to assist advertisers in finding the best placements for their brands across Alphabet's various platforms, while Video View seeks to maximize the number of views for video ads. Google reports that early tests indicate advertisers have seen a 40 percent increase in ad views thanks to these new tools.

Google Bard, Google's chatbot, was launched in all EU countries, including Sweden. The chatbot had several new updates, such as the ability to read its responses aloud to the user, accept image input, and understand more languages than before. When we asked Bard about the differences between ChatGPT and Bard, it described that Bard has access to a broader spectrum of information and is designed more as an information service. This means it's better at providing information than holding conversations, unlike ChatGPT.

Google has also partnered with Nvidia to predict extreme weather using AI. Predicting local weather has historically been challenging, but Google and Nvidia believe that AI can assist in this endeavor. Researchers at Nvidia have successfully developed a model that can predict hurricanes up to six weeks in advance.

Image: Midjourney/Hanna Berglind

There are rumors saying that Meta are also making significant investments in AI chatbots. According to the Financial Times, Meta plans to launch a series of AI chatbots with different personalities as early as September. These AI chatbots are intended to assist users with searches and recommendations. Meta's hope is that the chatbots will increase engagement among users. Internally at Meta, they have also begun testing AI tools, such as the AI chatbot "Metamate," which is designed to help employees summarize meetings and write code.

Apple, who's has been less vocal about AI in the public sphere compared to Google and Meta, is now rumored to be in the process of creating its own AI chatbot. It's still unclear how Apple's AI chatbot, Ajax, will be used, but there's speculation that it will compete with OpenAI's ChatGPT.

It's hard to discuss the tech giants' summer without mentioning Elon Musk's new AI startup, xAI, which he announced with a tweet in July. The purpose of xAI is described as "understanding the true nature of the universe," but what that entails remains unclear.

OpenAI's ChatGPT is being integrated into an increasing number of contexts. Car manufacturer Mercedes announced that ChatGPT will be integrated into the infotainment systems of their vehicles, something currently being tested in the U.S. The system is intended to provide suggestions for tourist attractions, for example. Meanwhile, travel company TUI has announced they will begin integrating ChatGPT into their app for user interactions. The AI is designed to generate personalized travel tips, among other things.

OpenAI launched its latest language model, GPT-4, in March 2023. During the summer, it was reported that OpenAI had applied for a trademark with the US Patent and Trademark Office for "GPT-5". While it's still too early to say how GPT-5 will differ from its predecessor, the trademark application suggests enhanced interactivity and a greater emphasis on voice-based features.

The expanding role of AI in journalism: nationally and globally

In July, the New York Times reported that Google is developing a new AI tool for journalism. The tool, named Genesis, is intended to function as an assistant in journalists' daily work, helping them write news articles.

Meanwhile, the American news agency AP reports that they are initiating a two-year collaboration project with OpenAI. It is reported that the parties will explore potential applications for generative AI in news products and services. OpenAI will also gain access to large amounts of text from AP to use as training data for their models.

Image: Midjourney/Hanna Berglind

This summer, Aftonbladet, the Swedish newspaper, gave the public an example of how AI can be applied to journalism. Aftonbladet used the language model GPT-4, the model behind the current ChatGPT, in their analysis of the Sweden Democrat-affiliated YouTube channel Riks. The AI identified over 10,000 different conversation topics, with the most common being the Sweden Democrats and the Social Democrats. The AI was able to see that the Social Democrats were often associated with words like desperate, power-hungry, and unreliable in Riks' YouTube clips, while the Sweden Democrats were instead described as demonized, misinterpreted, and misunderstood.

Swedish companies investing in AI

The Swedish salad farm, Kabbarps Garden, located outside of Malmö, is leveraging AI to monitor its salad production. The system analyzes the growth of salads using cameras in the greenhouse ceiling and sends alerts if the plants are stressed, lacking water, or infested with aphids. They hope that the system will increase production by 20%. However, using AI to monitor crops isn't new - the Swedish fresh herb producer, Svegro, has previously mentioned that they use AI to oversee and optimize their production.

Image: Midjourney/Hanna Berglind

Another Swedish company, Storytel, has announced plans to produce audiobooks using AI-generated voices. These AI voices are intended to both compete with and complement human narrators. Storytel's Head of Production, Johan Ståhle, states that production costs could be reduced by 95 percent due to the technology.

Earlier this year, Spotify launched its "AI DJ" in the USA, Canada, the UK, and Ireland. This summer, the AJ DJ was rolled out to 50 new markets, including Sweden. The AI DJ creates tailored song suggestions based on a user's listening patterns, paired with an AI voice that announces the artist's name and song title. The service is still in beta and will be refined over time.


Getting Ready for Fall

The summer has brought many news within the AI sphere, where generative AI, journalism, the tech giants' innovations, and Swedish companies' adoption of AI, have taken center stage. And these are just a few of all the news that has revolved around AI this summer.

Now, we are facing a fall where the pace of AI development is predicted to be faster than ever. Is your company prepared for the AI revolution? Or do you need assistance navigating the AI landscape?

Contact us at hello@violet.ai

About Violet

Violet AI was founded in 2018 as one of the first pure-play AI agencies in the Nordics. Today, Violet consists of a fast-growing consulting and advisory team and four subsidiaries with a total of 50 employees. We specialize in Machine Learning, Advanced Analytics, Intelligent Automation, System Development, and AI Strategy.