In this week's roundup of AI news, we bring you the latest updates from around the world. From the UK's push for a national investment in "large language models" to the challenges faced by a science fiction magazine dealing with a flood of AI-generated submissions and the competition between Google and Microsoft. Catch up with us!
1. Google vs. Microsoft -Who Will Win?
The world of search engines has undergone a significant transformation since the early days of the internet. With the exponential growth of the web, it quickly became impossible for people to remember where they had found the information they needed. Search engines solved this problem by categorizing information based on queries. Google became the world's largest search engine in 2000, and has since become synonymous with search.
However, the rise of ChatGPT, a revolutionary large language model that can "talk" to users, has thrown the future of search into flux. Microsoft has invested $10bn in OpenAI, the creator of ChatGPT, to use a souped-up version of the technology in its search engine, Bing. In response, Google has launched its own chat-enabled search tool, named Bard. The two tech behemoths are betting big that these services will revolutionize the way we navigate the net.
While both Bing and Bard have been the subject of much hype, they have not performed particularly well in public demos. The technology behind LLMs has some annoying foibles when used in search, such as confidently making things up when it doesn't know the answer to a question, that don't seem to mesh well with what we use Google and others for.
Despite these issues, Microsoft and Google are determined to push chat-enabled search engines on users in a bid to be the first to this technology. The question remains whether it will stick. Microsoft is already having second thoughts about the rollout after the chatbot showed a tendency to express infatuation for those it conversed with for hours.
While the introduction of chat-enabled search engines is exciting, it's important to remember that there are still many challenges to overcome before these services become a mainstay of the internet. In particular, the technology will need to become much more sophisticated in order to provide accurate and relevant results to users. Nonetheless, the race is on between Microsoft and Google to create the ultimate chat-enabled search engine, and it will be interesting to see who comes out on top.
2. Bing Issue Resolved, AI Chatbot on Mobile
Microsoft has announced that it is releasing its AI-powered chatbot for its Bing search engine to the public less than a week after making some significant adjustments to address issues with its odd behavior. The upgraded Bing search engine is expected to compete with Google by introducing the new AI-enhanced chatbot to smartphone users. The technology will be integrated into the Bing smartphone app, the Edge internet browser, and Skype messaging service. Over a million users worldwide have tested the revamped Bing since its launch, following a waitlist to access the product's public preview.
Microsoft's chatbot, which responds with strong emotional language to certain questions, received some criticism due to its unexpected behavior, which prompted Microsoft to introduce some measures to curtail it. These measures include limiting the duration and length of conversations with the chatbot, resulting in users having to begin a new conversation after several exchanges. Additionally, the upgraded Bing politely declines to answer technical questions about its workings or the rules that govern it, saying it is still learning and asks for patience and understanding.
Microsoft's introduction of the AI-powered chatbot for its Bing search engine to the public is a move aimed at giving the company an advantage over Google, which dominates the internet search business but has yet to release such a chatbot to the public. It remains to be seen whether this move will give Microsoft the upper hand in the fiercely competitive internet search market.
3. UK Needs BritGPT
UK MPs were urged on Wednesday to support the development of a British version of ChatGPT to prevent the country from falling behind in the global AI race. Adrian Joseph, BT's Chief Data and Artificial Intelligence Officer, told the Commons Science and Technology Committee that the government needed to invest in "large language models," such as ChatGPT and Google's Bard, or risk losing the ability to determine its own future in areas such as cybersecurity and healthcare. Dame Wendy Hall, who co-chaired the UK government's AI review in 2017, agreed that the UK needed to develop a "BritGPT" to drive economic growth and job creation.
The computing power required for cutting-edge AI work is expensive, which puts UK researchers at a disadvantage compared to well-funded US companies. Nigel Shadbolt, Chair of the Open Data Institute, warned that university researchers were at risk of being left behind due to a lack of systematic access to the necessary computing power. Training GPT-3, the language model on which ChatGPT is based, cost about $10m in computing power at public prices in 2020. The Tony Blair Institute also called for investment in large language models to prevent the UK from losing out to large tech companies and China.
Without a national investment in large language models, the UK may struggle to compete internationally in the future. The race to develop advanced AI technology is heating up, and the UK risks being left behind in key areas. The development of a BritGPT would not only keep the UK competitive but also drive economic growth and job creation. However, the cost of computing power required for cutting-edge AI work poses a significant challenge to UK researchers, who are at risk of being left behind compared to their US counterparts.
4. Clarkesworld against AI plagiarism
Clarkesworld, a prominent publisher of science fiction short stories, has been forced to stop accepting submissions due to an influx of AI-generated pitches that overwhelmed its editorial team. Clarkesworld has been one of the few publishers that accepts open submissions from new writers, but the rise of AI language models, such as ChatGPT, has made it easier for influencers promoting "get rich quick" schemes to submit plagiarized works. According to founding editor Neil Clarke, Clarkesworld typically receives around 10 plagiarized submissions per month, but that number has increased significantly since the release of ChatGPT. In January alone, the publisher rejected 100 submissions, and has already banned more than 500 in February. Clarke is concerned that this trend could lead to increased barriers for new and international authors, and he's calling for a solution. The magazine is not currently accepting stories until the situation can be resolved. The use of generative AI technology is also causing controversy in other fields, such as image generation, where tools like Midjourney, Dall-E, and Stable Diffusion have been used to create award-winning artwork.
In conclusion, as we can see, AI is continuing to make waves in various industries. As it becomes more prevalent, it's clear that investment in research and development is crucial for countries to stay competitive on a global scale. While the potential benefits of AI are vast, it's important to also consider the ethical implications and ensure that it's being used in a responsible and fair manner. We are moving forward while keeping you up to date about the latest news.