January News Recap: What’s Happening in the AI World?
The National Cyber Security Centre (NCSC) is warning about an expected increase in cybercrime that is facilitated by AI technology.
Artificial intelligence is being used to make the web more accessible for everyone, improving things like website navigation and content comprehension.”
Two months back, OpenAI introduced GPTs, a tool allowing users to personalize ChatGPT for specific tasks or roles. This feature is especially exciting for advanced users who prefer not to repeatedly teach ChatGPT the same things, and the best part is, you don’t need any coding skills to do it.
Initially, OpenAI promised a GPT Store where users could share and even make money from the GPTs they created. They’ve delivered on the first part by launching the GPT store. They’ve also pledged to start a revenue program for GPT builders sometime in Q1 2024.
OpenAI stated, “Initially, US builders will be compensated based on how much users engage with their GPTs. We’ll provide more details on payment criteria soon.” But does this mean builders outside the US won’t earn anything from this program right away? We’ll have to wait and see.
If you’re curious about creating your own GPT using the GPT Builder, check out our article: “How to create custom GPTs in ChatGPT.
Introducing the latest version of the ChatGPT Team plan, now accessible for you
So far, ChatGPT has had two payment options: ChatGPT Plus for individual users, and ChatGPT Enterprise for big companies like Zapier, PwC, and Canva. But what if you’re somewhere in between?
Well, OpenAI has introduced a new plan called ChatGPT Team this month. It gives you access to all the fancy features of ChatGPT Plus, but also provides a workspace where your team can collaborate, admin controls for managing your team, and the ability to have full control over your business data, just like with ChatGPT Enterprise.
The cost for ChatGPT Team is a bit more than ChatGPT Plus. It’s $25 per person per month, with a minimum of two users. But considering the extra features and privacy it offers, it’s still a pretty reasonable deal, especially for small to medium-sized businesses who can now access features similar to those of ChatGPT Enterprise.
OpenAI has introduced new embedding models and made updates to its API, while also reducing price
If you’re really into using OpenAI’s models, you’ll be happy to hear about some exciting updates this month. OpenAI launched two new embedding models, lowered prices for GPT-3.5 Turbo, and made it easier for developers to handle API keys and keep track of API usage. For more details about these new models and pricing changes, you can read the official announcement from OpenAI.
NCSC: AI will substantially increase cyber threats in next two years, only be offset by counter-use of AI
A recent report from the UK’s National Cyber Security Center (NCSC) suggests that artificial intelligence (AI) is likely to amplify both the frequency and impact of cyber attacks in the next couple of years. The report states that various kinds of cyber threat actors, whether they are from state or non-state entities, and regardless of their level of expertise, are already utilizing AI to some extent.
The NCSC highlights several areas where AI is expected to be employed by threat actors
- Social engineering attacks: AI can be used to create more convincing interactions with victims, such as crafting deceptive documents or messages, without the usual errors that might give away phishing attempts.
- Data analysis: AI can quickly sift through vast amounts of data to pinpoint valuable assets for theft or exploitation, making cyber attacks more impactful.
- Enhancing efficiency: AI can streamline various aspects of cyber attacks, including the construction of ransomware, reconnaissance activities, phishing schemes, and coding tasks.
- Malware development: AI holds the potential to generate malware that is capable of evading detection by existing security measures, particularly if it’s trained on high-quality exploit data. The NCSC warns that well-resourced entities, such as certain states, may possess extensive repositories of malware that could effectively train AI models for this purpose.
In essence, the report underscores the increasing role of AI in cyber attacks and the need for robust cybersecurity measures to counteract this evolving threat landscape.
The UK and Canada agreed to work together on AI computing.
Currently, many countries are striving to lead the way in the field of artificial intelligence (AI). The United Kingdom (UK) and Canada are now teaming up to work together towards this goal. They’ve signed an agreement to join forces on computing power to push forward AI research and development.
In simpler terms, to make progress in AI, you need both powerful computers and lots of data. Having access to advanced computing infrastructure helps countries compete on a global scale. Through this agreement, the UK and Canada will collaborate on key areas of research, such as biomedicine. And they’re not limiting themselves — they’re open to teaming up with other countries that share similar goals in sharing computing resources.
Aside from AI, the partnership between the UK and Canada also extends to other fields like quantum computing, semiconductors, and clean energy.
Scientists are working on creating artificial intelligence (AI) to help make the internet easier to use for everyone
For individuals with disabilities, using the internet can be really difficult. To help with this problem, researchers at The Ohio State University have been developing a smart computer program. This program can do complicated tasks on any website just by listening to simple voice commands.
Yu Su, one of the researchers involved, explains, “Some people, especially those with disabilities, find it hard to surf the web. We use computers more and more in our everyday lives, but there are many obstacles that make it hard for everyone to access them. This makes the gap between those who can easily use computers and those who can’t even wider.”
A robot has been taught how to read braille twice as fast as people
Scientists at the University of Cambridge have made a new kind of sensor that can understand braille faster than most people. They taught this sensor to recognize braille using smart computer programs. It can read at a speed of 315 words per minute and with 90% accuracy. Even though this sensor isn’t meant to help people with disabilities, it’s very good at feeling things delicately, which is important for reading braille. This could help in making robot hands or artificial limbs that can feel things like human fingertips do.
Conclusion
The National Cyber Security Centre (NCSC) warns of an impending rise in cybercrime facilitated by artificial intelligence (AI) technology. However, AI is not only a potential tool for cyber threats but also a solution to improving web accessibility for all users. OpenAI’s recent launch of the GPT Store introduces a platform for users to personalize ChatGPT, enhancing its utility and flexibility. Moreover, the introduction of the ChatGPT Team plan fills the gap between individual and enterprise users, catering to small to medium-sized businesses. OpenAI’s updates to its models and API further bolster its offerings to developers. On the global stage, partnerships like the one between the UK and Canada signify collaborative efforts to advance AI research and development. Meanwhile, initiatives such as the development of AI-powered programs for internet accessibility and the creation of robotic sensors capable of reading braille demonstrate the multifaceted applications of AI in addressing diverse challenges.
Follow 9 tools for understanding how software development grows in 2024 .
